The paper below is a draft version.

You can download the finally published version with the

correct pagination for free at:
http://clp.oxfordjournals.org/cgi/reprint/63.1.511?ijkey=2uHtwsXyKaQpBCO&keytype=ref

European Contract Law: Are No Oral Modification Clauses not

worth the paper they are written on?
DCFR II.-4:105 v CISG 29(2), UNIDROIT Principles 2.1.18, UCC § 2-209,
and Comparative Law

Florian Wagner-von Papp*

Mephistopheles: But one thing still: in case of
life or death, I pray /
Give me a written line or two. //
Faust: What, pedant! Something written do you
ask of me? /
Was neither man nor word of man yet known to
you?1

1. Introduction
After initial hesitations,2 the European Commission now seems determined to force the pace
in European contract law harmonization.3 It is still unclear whether the European Union is

*
This paper is based on a presentation delivered on 25 February 2010 in the Current Legal Problems lecture
series at the UCL Faculty of Laws in London. I would like to thank the UK Supreme Court Justice Lord
Mance both for chairing the session and for very helpful comments. Particular thanks are due to Mathias Siems
(UEA) for extensive comments on an earlier draft. For further comments, I would like to thank Stefan
Vogenauer (Oxford), Konstanze von Papp, and my colleagues at UCL George Letsas, Prince Saprai and
Ioannis Lianos, as well as members from the audience. While I would love to shift responsibility for remaining
errors to any or all of these persons, I am afraid I have to bear this responsibility alone. The reader finding such
errors is kindly asked to follow the suggestion in n 253.
1
Johann Wolfgang von Goethe, Faust – Eine Tragödie (Tübingen: Cotta 1808) 107: ‘Mephistopheles: Nur
eins! – um Lebens oder Sterbens willen, / Bitt’ ich mir ein Paar Zeilen aus. // Faust: Auch was geschriebnes
forderst du Pedant? / Hast du noch keinen Mann, nicht Mannes-Wort gekannt?’ (English translation by GM
Priest).
2
The European Parliament had called for a ‘European Civil Code’ as early as 1989. See European Parliament,
Resolution A2-157/89, OJ C 158/400 of 26 June 1989; European Parliament, Resolution A3-0329/94, OJ C
205/518 of 25 July 1994; Resolution B5-0228, 0229–0230/2000, OJ C 377/323, 326 [28] of 29 December
2000. The European Commission first responded – following a request by the European Council in Tampere
1999 – in its Communication of 11 July 2001 on European Contract Law, COM(2001)398 final; see also
European Commission, Communication of 12 February 2003, ‘A more coherent European contract law – an
1

Electronic copy available at: http://ssrn.com/abstract=1650501

heading for non-binding model rules, a binding or non-binding toolbox for the EU legislator,
a recommendation of a model law, an optional instrument as the so-called ‘28th system’,4 an
EU Directive on contract law, or – least likely of all – full harmonization by way of an EU
Regulation on Contract Law or ‘[t]he creation of a full-fledged European Civil Code,
replacing all national rules on contracts.’5 It is, however, increasingly likely that the EU is
going to take harmonizing action in some form in the area of contract law. The next step is
the drafting of a Common Frame of Reference, the adoption of which was initially
envisaged for 2009, was then postponed to 2010, and now is unlikely to be finalised before
2012.6
In order to aid the drafting of this ‘political’ Common Frame of Reference, the EU
commissioned and funded an academic study. 7 A host of preeminent comparative
researchers, organized in various groups, eventually published the ‘Draft Common Frame of
Reference’ (DCFR) containing principles, definitions and model rules of European private

action plan’, COM(2003)68 final; European Commission, Communication of 11 October 2004, ‘European
contract law and the revision of the acquis: the way forward’, COM(2004)651 final; European Commission,
Report of 23 September 2005, ‘First annual progress report on European contract law and the acquis review’,
COM(2005)456 final; European Commission, Report of 25 July 2007, ‘Second progress report on the
Common Frame of Reference’, COM(2007)447 final.
3
European Commission, ‘Green Paper on policy options for progress towards a European Contract Law for
consumers and businesses’, COM(2010)348 final, adopted 1 July 2010 (also cf Press Release IP/10/872 of the
same day); European Commission, ‘A Digital Agenda for Europe’, Communication of 19 May 2010,
COM(2010)245 final, 13 (‘The Commission will: . . . [p]ropose by 2012 an optional contract law instrument
complementing the Consumer Rights Directive to overcome the fragmentation of contract law, in particular as
regards the online environment’); European Commission, ‘Europe 2020 – A strategy for smart, sustainable and
inclusive growth’, Communication of 3 March 2010, COM(2010)2020 final, 21 (‘The Commission will
propose action to tackle bottlenecks in the single market by: . . . Making it easier and less costly for businesses
and consumers to conclude contracts with partners in other EU countries, notably by offering harmonised
solutions for consumer contracts, EU model contract clauses and by making progress towards an optional
European Contract Law . . .’).
4
See Press Release IP/10/872 of 1 July 2010, 2 and 3. Scots law is apparently not given any credit. The Green
Paper (n 3), 9, does not help either – it speaks of a ‘2nd Regime’ in each Member State.
5
Press Release IP/10/872 of 1 July 2010, 2, and Options 1-7 in the Green Paper (above n 3). Those preparing
the Draft Common Frame of Reference (see the text following this footnote) proceeded with their work behind
a veil of ignorance regarding the intended legal nature of the end product, cf C von Bar, ‘Coverage and
Structure of the Academic Common Frame of Reference’ [2007] European Review of Contract Law 350, 351–
2; H Beale, ‘The Future of the Common Frome of Reference’ [2007] European Review of Contract Law 257,
259. For a critique, see W Ernst, ‘Der “Common Frame of Reference” aus juristischer Sicht’ (2008) 208
Archiv für die civilistische Praxis 248, 257–78.
6
For the postponement until 2010, see European Commission, ‘Evaluation of the Hague Programme’,
Communication of 10 June 2009, SEC(2009)767 final, 115. Given that an Expert Group for the revision of the
academic Draft Common Frame of Reference (on this ‘DCFR’ see the text following this footnote) has been
set up on 26 April 2010, that 12 monthly meetings are planned (see the synthesis of the first meeting of the
Expert Group on 21 May 2010), that the Group’s mandate ends only on 26 April 2012 (Art 4(5) of the
Commission Decision 2010/233/EU, OJ L 105/109 of 27 April 2010), and that the European Commission
envisages an optional instrument for 2012 (Digital Agenda, n 3 above), it appears unlikely that the CFR will be
finalised before 2012.
7
cf European Commission, Action Plan, COM(2003)68 final [63] and European Commission, ‘The way
forward’, COM(2004)651 final, part 3.1.
2

Electronic copy available at: http://ssrn.com/abstract=1650501

law, first in an interim outline edition submitted to the Commission at the end of 2007,8 then
in an outline edition published early 2009,9 and finally in the full edition in six volumes
published in late 2009 and early 2010.10
While the mere academic and non-political nature of the DCFR is often
emphasized,11 the European Parliament has stated that the DCFR ‘should already be used as
a non-binding legal tool for European and national legislators’ and ‘insists that already now
the relevant provisions of the DCFR be given systematic and detailed consideration in all
forthcoming Commission proposals and impact assessments affecting contract law’. 12
Advocate General Trstenjak has referred to the DCFR in several of her opinions.13 Taking
into account that large-scale projects often develop a dynamic of their own, there is a
possibility that the choices the drafters of the DCFR have made are going to influence the
drafting of the political CFR or future instruments, not necessarily because they are
considered the best solution, but simply because they have already been made.14

8
C von Bar, E Clive, H Schulte-Nölke et al (eds), Principles, Definitions and Model Rules of European
Private Law – Draft Common Frame of Reference. Interim Outline Edition (Munich: Sellier European Law
Publishers, 2008).
9
C von Bar, E Clive, H Schulte-Nölke et al (eds), Principles, Definitions and Model Rules of European
Private Law – Draft Common Frame of Reference. Outline Edition (Munich: Sellier European Law Publishers,
2009), available at <http://ec.europa.eu/justice_home/fsj/civil/docs/dcfr_outline_edition_en.pdf>.
10
C von Bar, E Clive, H Schulte-Nölke et al (eds), Principles, Definitions and Model Rules of European
Private Law – Draft Common Frame of Reference. Full Edition (Munich: Sellier European Law Publishers,
2009, and Oxford: Oxford University Press, 2010) – Prepared by the Study Group on a European Civil Code
and the European Research Group on Existing EC Private Law (Acquis Group), based in part on a revised
version of the Principles of European Contract Law.
11
eg European Parliament, Resolution on the common frame of reference for European contract law, 3
September 2008, P6_TA(2008)0397, OJ C 295 E/31 of 4 December 2008, recital [D]; von Bar, Clive, Schulte-
Nölke et al, DCFR Outline Edition, Introduction [4], 6.
12
European Parliament, Resolution on the Communication from the Commission to the European Parliament
and the Council – An area of freedom, security and justice serving the citizen – Stockholm programme, 25
November 2009, P7_TA(2009)0090 [100].
13
eg Opinion 11 September 2008, case C-180/06 (Renate Ilsinger v Martin Dreschers) [2009] ECR I-3961
[49]–[52]; Opinion 18 February 2009, case C-489/07 (Pia Messner v Firma Stefan Krüger) (not yet reported in
ECR) [85]; Opinion 7 May 2009 case C-227/08 (Eva Martín Martín v EDP Editores, SL) (not yet reported in
ECR) [51]; Opinion 8 September 2009, case C-215/08 (E Friz GmbH v Carsten von der Heyden) (not yet
reported in ECR) in fn 62.
14
cf H Eidenmüller, F Faust, H C Grigoleit, N Jansen, G Wagner & R Zimmermann, ‘The Common Frame of
Reference for European Private Law – Policy Choices and Codification Problems’ (2008) 28 Oxford Journal
of Legal Studies 659, 668–9; House of Lords European Union Committee, ‘European Contract Law: the Draft
Common Frame of Reference’, 12th Report of Session 2008–9, HL Paper 95 (London: The Stationary Office,
2009) 28–9 [96]–[97]. While the costs for drafting the DCFR are sunk costs, and sunk costs ought to be
disregarded for future decisions, it is human nature not to disregard them: ‘We often see public and private
investment decisions made on the basis of what has already been invested’, H -B Schäfer & C Ott, The
Economic Analysis of Civil Law (trans M Braham, Cheltenham: Edward Elgar, 2004) 62. The Expert Group
set up on 26 April 2010 has the task to assist the Commission in ‘restructuring, revising and supplementing the
selected contents of the Draft Common Frame of Reference, taking also into consideration other research work
conducted in this area as well as the Union acquis’ (Art 2 of the Commission Decision 2010/233/EU, OJ
105/109, 110 of 27 April 2010). The solutions in the DCFR will thus certainly act as an anchor, and the extent
of any revision will depend entirely on the willingness of the Expert Group to reopen the discussion –
especially since ‘other research’ is only to be ‘taken into consideration’. Some commentators have doubted
3
 The efforts of the European Commission to harmonize the contract laws across the
European Union and in particular the academic DCFR have attracted both high praise and
severe criticism.15 It is not the purpose of this paper to embark on a discussion if or how

that the Expert Group in its current composition will be sufficiently critical of the DCFR, mentioning that ‘11
of the 14 professors’ on the Group appear to have a conflict of interest, because they have worked in the
CoPECL, the network contributing to the DCFR, and that only one of the professors has previously aired
criticism of the DCFR (J Basedow, H Eidenmüller, C Grigoleit, S Grundmann, N Jansen, E -M Kieninger, H -
P Mansel, W -H Roth, G Wagner & R Zimmermann, ‘Ein europäisches Vertragsrecht kommt – aber zu
welchem Preis?’, Frankfurter Allgemeine Zeitung no 149, 1 July 2010, 8). They also point out that the
timeframe for the Expert Group (above n 6) hardly leaves them time for a thorough re-evaluation of the DCFR
rules.
15
For a critical view see Eidenmüller et al, ‘Common Frame of Reference’ (2008) 28 Oxford Journal of Legal
Studies 659–708; S Whittaker, The ‘Draft Common Frame of Reference’ – An Assessment, commissioned by
the Ministry of Justice, United Kingdom (2008); House of Lords European Union Committee, ‘European
Contract Law’; J Mance, ‘Is Europe aiming to civilise the common law?’ [2007] European Business Law
Review 77, eg 97; idem, The Common Frame of Reference, (2010) Zeitschrift für Europäisches Privatrecht,
lead article in issue 3; R Zimmermann, ‘The Present State of European Private Law’ 57 American Journal of
Comparative Law 479, 490–503, 510–2 (2009), with numerous further references, esp to German and French
critics; Ernst, ‘Common Frame of Reference’ (2008) 208 Archiv für die civilistische Praxis 248; S Grundmann,
‘The Structure of the DCFR – Which Approach for Today’s Contract Law?’ [2008] European Review of
Contract Law 225; see also the contributions in HW Micklitz & F Cafaggi (eds), European Private Law after
the Common Frame of Reference (Cheltenham: Edward Elgar, 2010); for an extensive critique of the drafting
process of the DCFR, see Chapter 4 of Lucinda Miller’s PhD thesis, to be published with Oxford University
Press in 2011; for a summary of the Italian discussion, see B Pasa, ‘The DCFR, the ACQP and the Reactions
of Italian Legal Scholars’ [2010] European Review of Private Law 227 (noting, at 230, that ‘the commentaries
criticizing the DCFR and . . . the ACQP outnumber the favourable ones’); see further Study Group on Social
Justice in European Private Law, ‘Social Justice in European Contract Law: a Manifesto’ (2004) 10 European
Law Journal 653; JM Smits, ‘The Draft-Common Frame of Reference, Methodological Nationalism and the
Way Forward’ [2008] European Review of Contract Law 270; R Schulze & T Wilhelmsson, ‘From the Draft
Common Frame of Reference towards European Contract Law Rules’ [2008] European Review of Contract
Law 154. See also N Jansen & R Zimmermann, ‘Was ist und wozu der DCFR?’ [2009] Neue Juristische
Wochenschrift 3401; J Basedow et al, ‘Europäisches Vertragsrecht’, Frankfurter Allgemeine Zeitung no 149, 1
July 2010, 8; H Eidenmüller, ‘Party Autonomy, Distributive Justice and the Conclusion of Contracts in the
DCFR’, [2009] European Review of Contract Law 109 (draft available at <http://ssrn.com/abstract=1334648>);
JP Schmidt, ‘Der “juridical act” im DCFR’ [2010] Zeitschrift für Europäisches Privatrecht 304; further
references to critical (German) views can be found in fn 1 of E Hondius ‘England and the European Civil
Code’ [2008] European Review of Private Law 85. From a law & economics perspective, see the contributions
in G Wagner (ed), The Common Frame of Reference: A View from Law & Economics (Munich: Sellier
European Law Publishers, 2009) and in P Larouche & F Chirico (eds), Economic Analysis of the DCFR
(Munich: Sellier European Law Publishers, 2010). For positive views, see eg the EP Resolution
P7_TA(2009)0090; Beale, ‘Future’ [2007] European Review of Contract Law 257; H Schulte-Nölke, ‘EC Law
on the Formation of Contract – from the Common Frame of Reference to the “Blue Button” ’ [2007] European
Review of Contract Law 332; M Hesselink, ‘Common Frame of Reference & Social Justice’ [2008] European
Review of Contract Law 248; S Leible ‘Europäisches Privatrecht am Scheideweg’ [2008] Neue Juristische
Wochenschrift 2558; for a response to some of the criticism, see H Schulte-Nölke, ‘Arbeiten an einem
europäischen Vertragsrecht – Fakten und populäre Irrtümer’ [2009] Neue Juristische Wochenschrift 2161. For
further comments, see eg the contributions in issue 4 of the European Review of Private Law 2009.
For several reasons, one has to be careful not to mistake the sheer quantity of the critical commentary
as a conclusive sign for systematic weaknesses in the DCFR: First, there is an obvious sampling bias because
of the self-selection of commentators: those dissatisfied with some (major or minor) aspect of the DCFR have
a reason to publish their criticism, while those satisfied could still be in the silent majority. Secondly, a drafter
choosing a compromise solution will be attacked from both sides: It will, for example, be difficult to draft
rules that satisfy both hard-core law & economics scholars of the Chicago School and consumer protectionists
who join Otto von Gierke in calling for ‘a drop of social oil’ (or, considering the difference between the draft
Civil Code on which Otto von Gierke was commenting and the Consumer Acquis, a barrel of it). And, of
course, even comparative law scholars from the different Member States may have a certain bias towards the
solutions found in their home jurisdictions and may criticize deviations. Thirdly, the criticism differs vastly in
4
contract law should be harmonized across Europe. Nor does this paper seek to assess the
scholarly quality of the DCFR as a whole, or its usefulness for the purpose for which it was
drafted.16
The aim and scope of this paper is much more limited: It singles out one model rule
from the DCFR, namely DCFR II.-4:105, and assesses this rule, along with the
accompanying comments and notes, from a comparative perspective. DCFR II.-4:105 deals
with so-called ‘no-oral-modification’ (or ‘no variation except in writing’) clauses. Before
embarking on a comparative analysis, some background information on form requirements
and ‘no-oral-modification’ clauses may be useful.
Sam Goldwyn famously said that ‘an oral contract is not worth the paper it is written
on.’ Nice as the witticism is: while legal systems historically started out by enforcing only
formalized contracts,17 parties are today as a general rule free to form legally enforceable

intensity, extent, and direction. In part, the uncertainty about the intended end result (see n 5 above)
contributes to the amount of criticism: Nobody knows whether one has to argue against a non-binding toolbox
or against a European Civil Code or something in between. Even if one knew that the CFR would be used
‘only as a toolbox’, it would still be unclear what the toolbox metaphor entails (mere ‘background
information’ or a blueprint to be copy-pasted into future legislation?). As one distinguished commentator
remarked, ‘anything faintly resembling a coherent set of rules cannot . . . be a toolbox. Another reason [scil.
for considering the toolbox metaphor the ‘ugliest phrase of the year’] probably stems from my experience of
academic households, where toolboxes are often lavishly equipped . . ., with no one around who knows how to
employ the items competently.’ (CC Wendehorst, quoted in [2009] Zeitschrift für Europäisches Privatrecht
516). Furthermore, it is unclear whether the DCFR was meant to compile what the laws of the Member States
already provide (backward-looking perspective) or whether the DCFR should be aspirational, looking for the
best solutions (forward-looking perspective). Hence, where the drafters chose a backward-looking perspective,
they can be criticized for making the same mistake as the drafters of the German Civil Code at the end of the
nineteenth century, and where they were aspirational, they can be criticized for making political choices not
supported by comparative observations.
16
On the last point, see House of Lords European Union Committee, ‘European Contract Law’, 16 [38]:
‘[T]he approach adopted and choices made appear vulnerable to suggestions that they will not be beneficial or
usable in a practical sense.’; see also ibid, 24–5 [74]–[81].
17
See for the ancient Roman law, R Zimmermann, The Law of Obligations (Oxford: Oxford University Press,
1990/1996) 82–3. Even after the early sacral formalism of the pontifices had been overcome, formalism
remained the prevailing rule: the stipulatio was a strictly formalised, though initially oral act (Ulp Dig
45.1.1.pr.); later, a writing evidencing that the stipulatio had taken place relaxed strict compliance with some
of the painstaking requirements imposed earlier, see eg C 8, 37, 1 (Sev et Ant); Paulus Dig 24, 1, 57; 45, 1,
134, 2; C 8, 37, 10 (Leo); Inst III, 19, 12; C 8, 37, 14, 2 (Iust); see also Inst III, 21. De facto, the written form
thus became a substitute to the earlier strict oral form requirements, see eg, Zimmermann, Law of Obligations,
68–71, 79–82; B Nicholas, ‘The Form of the Stipulation in Roman Law’ (1953) 69 LQR 63-4 (Part I) and esp
233, 235–52 (Part II); see also J Gordley, Foundations of Private Law (Oxford: Oxford University Press, 2006)
290–1.
In England, the action of covenant was limited to agreements embodied in ‘specialties’, ie documents
under seal, at least from the late thirteenth and early fourteenth century onwards, see D Ibbetson, A Historical
Introduction To The Law Of Obligations (Oxford: Oxford University Press, 1999) 24–8; AWB Simpson, A
History of the Common Law of Contract (Oxford: Oxford University Press, 1975) 10–1 (noting that ‘there
never seems to have been a time when the royal courts regularly allowed actions to be taken on parole
covenants’) and 22. Debt or detinue based on informal agreements only covered actions for a certain sum or
specific goods (Ibbetson, ibid, 31–8; for services see Simpson, ibid, 148–52), even though creative drafting
allowed the use of debt based on a conditional bond as a functional equivalent to an action of covenant
(Ibbetson, ibid, 28–30; but this brought with it the possibility of wager). Subsequently, however, the
enforceability of informal contracts was recognized, based on trespass, and then trespass on the case in the
5
contracts informally, eg orally or by conduct,18 unless statutory law explicitly establishes a
form requirement.19, 20 And the same is true for contract modifications. Unless there is a

form of assumpsit (Ibbetson, ibid, 30–56, 95–6, 126–51; Simpson, ibid, 199–315), although there is some
uncertainty as to whether commonly practiced formalities, such as the handing over of a God’s penny, a
handshake, or a drink together, may at some point have been considered necessary ‘form’ requirements
(Ibbetson, ibid, 74–6). See also J Beatson, Anson’s Law of Contract (28th edn, Oxford: Oxford University
Press, 2002) 11–16; EA Farnsworth, Contracts (4th edn, New York, NY: Aspen, 2004) §§ 1.5–1.6 and 2.16;
and generally PS Atiyah, The Rise and Fall of Freedom of Contract (Oxford: Oxford University Press, 1979).
In nineteenth-century Germany, the possibility of informal contracting was nearly everywhere the
general rule, but in Prussia the General Land Law established form requirements for so many transactions that
the de facto general rule was said to be one of formalism, replacing the de iure general rule of informalism in
prALR I 4 § 94, I 5 § 109. Cf the explanatory notes of the First Commission in the drafting of the German
Civil Code, Motive zum Allgemeinen Theile, Vol 1 (Berlin: Guttentag 1888) 178–81, reprinted in B Mugdan,
Die gesammten Materialien zum Bürgerlichen Gesetzbuch für das Deutsche Reich Vol 1 (Berlin: Decker 1899)
450–1.
18
Throughout this paper, I will use the term ‘informal’ to denote everything that does not satisfy the form
requirement in question. Where ‘writing’ is required, this includes both oral modifications and modifications
by conduct; depending on the legal requirements of the jurisdiction, it may also include written statements that
lack signatures etc. Where a deed is required, even a written statement falling short of the requirements for a
deed would be ‘informal’ under this definition.
19
See generally K Zweigert & H Kötz, An Introduction to Comparative Law (3rd edn, Oxford: Oxford
University Press, 1998) 366: ‘In all modern legal systems there are some transactions which have to be in a
special form . . . Such formal requirements are, however, always exception to a general principle of
informality’; Zimmermann, Law of Obligations, 85–6; J Basedow, ‘Freedom of Contract in the European
Union’, [2008] European Review of Private Law 901, 906–7 (identifying freedom of form as one of the
aspects of freedom of contract).
Under Art 11 of the Convention on Contracts for the International Sale of Goods (CISG), a ‘contract
of sale need not be concluded in or evidenced by writing and is not subject to any other requirement as to form.
It may be proved by any means, including witnesses.’ (Note, however, the possibility of CISG Art 96
declarations, made, eg, by Argentina, Armenia, Belarus, Chile, Estonia, Hungary, Latvia, Lithuania, Paraguay,
the Russian Federation, and the Ukraine; China has made a similar declaration, stating that it does not consider
itself bound by CISG Art 11 and provisions relating to this Article).
Similarly, Art 1.2 of the UNIDROIT Principles provides that ‘[n]othing in these Principles requires a
contract, statement or any other act to be made in or evidenced by a particular form. It may be proved by any
means, including witnesses.’
The DCFR provides in II.-1:106(1) that a ‘contract or other juridical act need not be concluded, made
or evidenced in writing nor is it subject to any other requirement as to form.’ The Acquis Principles state in
Art 1:304 (Freedom of Form): ‘Unless provided otherwise, no form needs to be observed in legal dealings.’
(Research Group on the Existing EC Private Law (Acquis Group), Contract II (Munich: Sellier European Law
Publishers, 2009) 75–80).
For the US, see eg Restatement (Second) of Contracts § 4 (1981): ‘A promise may be stated in words
either oral or written, or may be inferred wholly or partly from conduct.’
For English law, cf Beckham v Drake (1841) 9 M&W 79, 92–3: ‘[T]he law makes no distinction in
contracts, except between contracts which are and contracts which are not under seal. I recollect one of the
most learned judges who ever sat upon this or any other bench being very angry when a distinction was
attempted to be taken between parol and written contracts, and saying “they are all parol unless under seal.” . . .
A contract not under seal is open to all the common-law requirements and incidents of a contract, whether in
writing or not.’ (Lord Abinger, CB). While the quoted passage argued that the position of a party to a contract
in writing should be no worse off than a party to an oral contract (specifically, the naming of only two of the
defendants in the written document should not exclude the liability of a third party as a principal), the
argument is clearly premised on the equivalence of oral and written contracts. On this point, the Exchequer
Chamber agreed, (1843) 11 M&W 315, 317 (Lord Denman, CJ), and in the House of Lords, the plaintiff
abandoned the argument that a distinction should be made, (1849) 9 ER 1213, 1215; for the principle of
informality in English law, see also SJ Whittaker in Chitty on Contracts Vol I (30th edn, London: Sweet &
Maxwell, 2008) para 4-001; E Peel, Treitel The Law of Contract (12th edn, London: Sweet & Maxwell, 2007)
para 5-003; also cf E Peden & JW Carter, ‘Entire Agreement – and Similar - Clauses’ (2006) 22 Journal of
Contract Law 1 (‘It seems trite to say that a contract may be written, oral , partly written or implied from
conduct.’).
6
statutory form requirement, contracts can be modified informally; today, this is true even
where the contract to be modified is in writing or in a deed.21

The Swiss law of obligations provides in Art 11(1) that ‘contracts, in order to be valid, are only
required to be in a particular form if prescribed by the law.’ Similarly, § 883 of the Austrian ABGB states: ‘A
contract can be concluded in writing or orally . . . These differences in form do not make a difference as
regards the obligation, except in the cases where the law provides otherwise.’ In Germany, the drafters of the
BGB had at first considered including a similar provision (see First Commission, 4th session of 22 September
1876, Protokolle I 150–1, reprinted in HH Jakobs & W Schubert, Die Beratung des Bürgerlichen
Gesetzbuches, Vol 2/1 (Berlin: Walter de Gruyter, 1985) 653; for the explanatory notes to the First Draft, see
Motive zum Allgemeinen Theile, Vol 1, 178–81, reprinted in Mugdan, Die gesammten Materialien Vol 1, 450–
1). Eventually, it was decided, however, that this principle became clear e contrario from the first sentence of
what is today § 125, which states that ‘a legal transaction which lacks the form that is required by law is
invalid’, see Vorkommission des Reichsjustizamtes, 14th session of 25 January 1891, Protokolle RJA 59,
reprinted in Jakobs & Schubert, Beratung, Vol 2/1, 662; see also BS Markesinis, H Unberath & A Johnston,
The German Law of Contract (2nd edn, Oxford: Hart, 2006) 81; C Hertel in J von Staudingers Kommentar
zum Bürgerlichen Gesetzbuch (Berlin: Sellier-de Gruyter 2004) § 125 para 3; K Larenz & M Wolf,
Allgemeiner Teil des Bürgerlichen Rechts (9th edn, Munich: CH Beck, 2004) § 27 para 3; J Ellenberger in
Palandt Bürgerliches Gesetzbuch (69th edn, Munich: CH Beck, 2010) § 125 BGB para 1.
For French law see B Nicholas, The French Law of Contract (Oxford: Oxford University Press, 1992)
59–61; S Whittaker in J Bell, S Boyron & S Whittaker, Principles of French law (2nd edn, Oxford: Oxford
University Press, 2008) 325–6.
For Japanese Law, see H Oda, Japanese Law (3rd edn, Oxford: Oxford University Press, 2009) 150–2.
Even Russian law, which used to have very extensive form requirements, has come to give effect to oral
juridical acts and those by conduct; nevertheless, there are still form requirements for a wide range of
transactions. See Art 158–9 of the Russian Civil Code (English translation in WE Butler, Civil Code of the
Russian Federation (Oxford University Press, 2002) 71–2); H Oda, Russian Commercial Law (2nd edn,
Leyden: Kluwer 2007) 86–7.
20
Of course, Goldwyn’s quote may have been referring to: (1) the fact that under US law, many contracts are
within a Statute of Frauds (cf Restatement (Second) of Contracts § 110, Uniform Commercial Code (UCC) §
2-201 (2003); EA Farnsworth, Contracts, § 6.2), and/or (2) the evidential question of proof.
If this is what Goldwyn meant, his witticism comes closer to the truth. Note, however, that even where a
contract is within some version of a Statute of Frauds (for the US, see above; for France, see Art 1341–6 of the
Code Civil), oral contracts may still be valid and enforceable: in the US, a written memorandum evidencing
the existence of the contract will often satisfy the Statute, even if it falls far short of a written contract,
Farnsworth, Contracts, § 6.7. In France, the expansive interpretation of ‘commencement de preuve’ in
Art 1347 and other exceptions, eg Art 1348, similarly make oral agreements enforceable as long as there is
something in writing (cf Nicholas, French Law, 61; D Tallon, ‘Contract Law’ in GA Bermann & E Picard
(eds), Introduction to French Law (Austin: Wolters Kluwer, 2008) 220–1 Zweigert & Kötz, Introduction, 369–
71); for merchants, there is a general exception in the Code de Commerce L110-3.
21
For English law, see Berry v Berry [1929] 2 KB 316: A written modification of a promise in a deed was held
to be valid, because equity, which prevails since the Judicature Act 1873, allowed parol agreements not to
enforce – or to enforce only with modifications – a deed. While even in equity the deed was not varied by the
parol agreement (Nash v Armstrong (1861) 10 Common Bench Reports (New Series) 259, 266–7, 142 ER 451,
455: ‘the bond is not released’), the promisee could not enforce it without incurring liability (ibid: ‘whereas
before he had a right to sue upon the deed, if he now exercises that right he renders himself liable to an action’).
In contrast, the position at common law was that a deed could only be discharged or modified by a deed, eg
West v Blakeway (1841) 2 Manning & Granger 729, 751–2, 133 ER 940, 950 (CP): ‘Now it is a well-known
rule of law, that Unumquodque ligamen dissolvitur, eodem ligamine quo et ligatur. This is so well established,
that it appears to me unnecessary to refer to cases. I will mention only Rogers v Payne. . . [T]he contract into
which he has entered can be discharged only by an instrument of the same nature as that by which the contract
was created.’ (Tindal, CJ); and in the same case, Bosanquet, J, stated: ‘No rule of law is more fully established
than this, that a contract under seal cannot be varied by a parol contract.’ See also Kaye v Waghorn (1809) 1
Taunt. 428, 429–30, 127 ER 900, 901 (CP); see, however, the qualification in Blake's Case, (1605) 6 Coke
Reports 43b-44a, 77 ER 318, 320 (CP), that this only applies insofar as the duty accrues from the deed with
certainty. See also Simpson, History of the Common Law, 25–7 (covenant); Ibbetson, Historical Introduction,
84; EG McKendrick in Chitty on Contracts § 22-030. Where statutory law requires the modification, or the
resulting contract, to be in writing, a parol modification will not be valid, Goss v Lord Nugent (1833) 5
Barnewall and Adolphus 58, 110 ER 713 (KB), although a parol attempt at modification may still be valid as a
7
 In commercial practice, however, the parties to a contract often want to exclude the
possibility of subsequent informal modifications. One of the reasons for such a preference
for formality is the desire to avoid disputes about the existence or content of informal
modifications, which could arise from misunderstandings, poor recollection of witnesses, or
from deliberately false testimony. Many contracts contain a clause such as:

No modification or alteration of this Agreement shall be validly made unless in

writing and signed by or on behalf of both parties hereto.

Such a clause is called ‘no variation except in writing’ clause or ‘no-oral-modification’

clause, often shortened to the ugly, but concise acronym ‘NOM-clause’.22 Theoretically, the
clause could provide for a form more exacting than writing, such as a deed or a notarial
authentification, but this would make the process of modification cumbersome and
consequently appears to be extremely rare in practice; alternatively, and more frequently,
the parties can agree on a form less exacting than writing. 23 It is not only ‘oral’

rescission (Morris v Baron [1918] AC 1 (HL)) or act as a waiver (Besseler Waechter Glober and Co v South
Derwent Coal [1938] 1 KB 408, 416–8; McKendrick in Chitty on Contracts § 22-041). For the problems
arising from the tenuous distinction between modification and rescission/novation, see below 6; for the
distinction between variation and waiver below 8.
In the US, even where the contract to be modified is within a Statute of Frauds, this in itself does not
necessarily prevent oral modifications, ‘as long as no essential terms were modified’, Farnsworth, Contracts
§ 6.2, or where the contract as modified is no longer within the Statute, Liebermann v Templar Motor, 236 NY
139, 140 NE 222 (1923). Before the seal was abolished, promises under seal could usually not be varied by
parol, cf for New York: Cammack v JB Slattery & Brothers, 241 NY 39, 45–6, 148 NE 781 (NY 1925).
In Germany, informal modifications of written contracts are valid, unless the original contract is
subject to a statutory form requirement or the content of the modification itself is such as to trigger a statutory
form requirement (see C Grüneberg in Palandt § 311 BGB paras 4, 5).
In France, oral modifications to a contract which exceeds the threshold of Art 1341 Code Civil cannot,
in principle, be proven by witness testimony, because Art 1341 Code Civil excludes such testimony for
statements not only before and at the time of the drafting of the document, but also for statements after the
drafting. It has to be borne in mind, however, that the expansive interpretation of ‘commencement de preuve’
in Art 1347 Code Civil and the exception for commercial contracts limit the practical importance of Art 1341
Code Civil (see above, n 20).
22
I will use ‘modify’ and ‘vary’ (and the corresponding nouns) indiscriminately. Given the international usage,
I tend to use ‘modify’ rather than ‘vary’, although in England and Wales there is a preference for the term
‘vary’. The German, Austrian, and Swiss equivalent to NOM-clases is ‘Schriftformklausel’ (literally: ‘written
form clause’).
23
The DCFR uses the more precise, but slightly awkward phrase ‘modification in certain form only’ (DCFR
II.-4:105) to allow for form requirements other than writing (in contrast to its predecessor, Art 2:106 of the
Principles or European Contract Law (PECL), which used the term ‘writing’). Similarly, Art 2.1.18 of the
UNIDROIT Principles now uses the phrase ‘modification in a particular form’ (in contrast to its predecessor,
Art 2.18 of the UNIDROIT Principles 1994, cf S Eiselen, ‘Modification or termination of contract and
formalities: Remarks on the manner in which the UNIDROIT Principles of International Commercial
Contracts may be used to interpret or supplement Article 29 of the CISG’ in: J Felemegas (ed), An
International Approach to the Interpretation of the United Nations Convention on Contracts for the
International Sale of Goods (1980) as uniform sales law (Cambridge: Cambridge University Press, 2007) 166
comment j). Conceptually, there are few limits to the creativity parties could unfold in establishing form
8
modifications that the clause wants to exclude, but also modifications by conduct. For the
sake of brevity, this paper will use the commonly employed acronym ‘NOM-clause’, fully
accepting that the term is slightly imprecise.
Given the widespread use of NOM-clauses in commercial practice, it seems curious
that their legal effect in many jurisdictions is far from clear. In some jurisdictions, courts
consider an informal modification of a contract to be valid, even where the contract contains
a NOM-clause. 24 Other jurisdictions and legal instruments enforce NOM-clauses and
consequently consider a subsequent informal modification to be invalid;25 but even among
these jurisdictions, there are variations as to the preconditions under and the extent to which
NOM-clauses are enforced.26
In light of the commercial demand for such clauses, instruments for legal
harmonization have generally come down in favour of the enforceability of NOM-clauses:
The US American Uniform Commercial Code (‘UCC’), the UNIDROIT-Principles of
International Commercial Contracts 2004 (‘UNIDROIT Principles’), and the Convention on
Contracts for the International Sale of Goods (‘CISG’) have all opted for their
enforceability.27
In contrast, DCFR II.-4:105 provides:

(1) A term in a contract requiring any agreement to modify its terms, or

to terminate the relationship resulting from it, to be in a certain form
establishes only a presumption that any such agreement is not intended to be
legally binding unless it is in that form.

requirements. See the conversation between Faust and Mephistopheles following the request for a contract in
writing (above n 1, at 107–8): ‘Faust: . . . What, Evil Spirit, do you want of me? / Brass, marble, parchment,
paper? Name it then! / Am I to write with graver, chisel, pen? . . . // Mephistopheles. . . . Just any little sheet
will do, it's all the same. / With one wee drop of blood you sign your name.’ (‘Faust: . . . Was willst du böser
Geist von mir? / Erz, Marmor, Pergament, Papier? / Soll ich mit Griffel, Meißel, Feder schreiben? . . .
Mephistopheles: . . . Ist doch ein jedes Blättchen gut. / Du unterzeichnest dich mit einem Tröpfchen Blut’). See
also Zimmermann, Law of Obligations, 86 in fn 115: ‘[I]n early Bavaria and Alemannia transfers of land had
to be performed in the presence of a certain number of small boys, who, after attending the ceremony, were
treated to a box on the ear in order that they might keep a vivid remembrance of what had happened. Without
such a box, the transfer was void.’ (citing P Vinogradoff, Outlines of Historical Jurisprudence Vol I (1920)
364). While beating little boys has gone out of fashion for good reasons, the handshake or drink together
common in medieval England (n 17) could be a socially acceptable alternative. In practice, however, few
commercial parties opt for such more interesting form requirements, although one Austrian court considered
an agreement sealed by a handshake to overcome a NOM-clause, see Oberlandesgericht Innsbruck 18
December 2007, case 1 R 273/07t, <http://cisgw3.law.pace.edu/cases/071218a3.html> (Steel bars case).
24
Below 2.
25
Below 3.
26
See below 3, 5–9.
27
UCC § 2-209(2); UNIDROIT Principles Art 2.1.18; CISG Art 29(2); see below 3.
9
 (2) A party may by statements or conduct be precluded from asserting
such a term to the extent that the other party has reasonably relied on such
statements or conduct.

While it may not be obvious at first glance, 28 a closer analysis of the rule and the
accompanying comments shows that this model rule denies NOM-clauses any effect.29
Part 2 of this paper describes several jurisdictions that treat NOM-clauses as
unenforceable, or did so in the past. Part 3 describes jurisdictions and legal instruments that
give effect to NOM-clauses; in many jurisdictions, this approach has overcome earlier
resistance. Part 4 discusses the policy reasons counselling for and against the enforcement
of NOM-clauses, and concludes that NOM-clauses should generally be enforceable. Parts
5–9 address several limitations to the enforceability of NOM-clauses in jurisdictions that
usually do consider them to be effective. Part 10 critically analyses the solution adopted by
the drafters of the DCFR in DCFR II.-4:105.30 With the greatest possible respect for the
drafters of the DCFR, I conclude that the comparative analysis regrettably opted for breadth
instead of depth. The analysis is devoid of any discussion of policy arguments. The
comparative analysis of the national solutions is mostly based on ‘the law in the books’,

28
M del Pilar Perales Viscasillas, ‘Modification and termination of the contract (Art. 29 CISG)’, 25 Journal of
Law and Commerce 167, 176–7 (2005) assumes that Art 2:106 of the Principles of European Contract Law
(PECL), ie the predecessor to DCFR II.-4:105 whose wording and comments are nearly identical with the
DCFR rule, is ‘similar’ to CISG Art 29(2) and UNIDROIT Principles Art 2.1.18. With regard to DCFR II.-
4:105, MJ Bonell & R Peleggi diplomatically note – as the first of ‘[t]he most significant divergences of a
technical nature’ – that, in contrast to Art 2.1.18 of the UNIDROIT Principles, the DCFR merely establishes a
presumption that the contract containing a NOM-clause may not be otherwise modified, without enquiring into
the effect of the ‘presumption’ established (‘UNIDROIT Principles of International Commercial Contracts and
Draft Common Frame of Reference: a synoptical table’, (2009) 14 Unif L Rev ns 437, 443). S Eiselen, Guide
to Article 29 – Comparison with PECL, <http://cisgw3.law.pace.edu/cisg/text/peclcomp29.html> (2002),
comment f, correctly recognizes the crucial difference that ‘[a]n informal agreement to modify or terminate a
contract may . . . be interpreted as “an implied abrogation” of the clause itself [scil.: under PECL Art. 2:106]. . .
whereas this is not the case with the CISG.’ The clear divergence between the DCFR and the CISG on this
point is also recognized in S Troiano, ‘CISG’s Impact on EU Legislation’ [2008] Internationales Handelsrecht
221, 241–2.
29
Below 10.
30
I should point out that the rule, the comments and the comparative notes of DCFR II.-4:105 are nearly
identical to its predecessor, Art 2:106 of the Principles of European Contract Law (PECL). There is some, but
very little evidence of updating of the PECL rule, comments and notes. Because of the potential political
impact of the DCFR and the greater resources available to the DCFR drafters, I focus on the DCFR rather than
the PECL. My objection is against ‘the uncritical incorporation of the PECL’ (see Eidenmüller et al, ‘Common
Frame of Reference’ (2008) 28 Oxford Journal of Legal Studies 659, 699). The Principes Contractuels
Communs of the Association Henri Capitant also uncritically accept the PECL rule; they suggest only
changing the positioning of the rule from the chapter on formation to the chapter on the content of the contract.
See Association Henri Capitant des amis de la culture juridique française & Société de législation comparée
(eds), Projet de Cadre Commun de Référence – Principes Contractuels Communs (Paris: Société de
Législation Comparée, 2008, with a preface by B Fauvarque-Cosson & D Mazeaud and an introduction b G
Wicker & J-B Racine) 232-4.
10
contrary to the principle of comparative law to focus on the ‘law in action’, and the
description of the national solutions is not always accurate. What is more, the model rule
itself is internally inconsistent. I conclude by asking whether the deficiencies in DCFR II.-
4:105 are in any way representative for the DCFR and what policy questions should be
addressed when drafting a rule on NOM-clauses.

2. The Freedom of Contract Argument for not Enforcing NOM-clauses: ‘The

Clause Which Forbids a Change May Be Changed Like Any Other’
In several jurisdictions, courts have decided that the presence of a NOM-clause in a contract
cannot prevent a subsequent oral modification from becoming effective.
The underlying line of thought can be summarized as follows: (i) Oral contracts are
as binding as written contracts. (ii) An oral modification at least implicitly derogates from
the NOM-clause. (iii) The NOM-clause and the agreement to derogate from the original
contract are of equal weight; the NOM-clause does not have the ‘quasi-constitutional’ effect
of introducing a writing requirement for subsequent modifications where none existed
before. (iv) Therefore, the oral modification is a binding contract, which varies the original
written contract and supersedes the NOM-clause. The principle of freedom of contract
demands that the subsequent oral modification be given effect,31 because two parties have
agreed to the modification and wanted it to be binding – or so the argument goes.32

United States
The language used by the courts to justify the unenforceability of NOM-clauses is
remarkably similar across many jurisdictions. In the New York case Beatty v Guggenheim
Exploration, Judge Benjamin Cardozo used the following words:
Those who make a contract may unmake it. The clause which forbids a change may
be changed like any other. The prohibition of oral waiver may itself be waived.

31
Always provided that all other requirements for a binding contract are present, in particular consideration
where required.
32
See the decisions described in the text below. Also cf W Flume, Allgemeiner Teil des Bürgerlichen Rechts II
– Das Rechtsgeschäft (4th edn, Berlin: Springer, 1992) 264–5; D Reinicke, ‘Die Bedeutung der
Schriftformklausel unter Kaufleuten’ [1976] Der Betrieb 2289, 2290; L Häsmeyer, ‘Die Bedeutung der Form
im Privatrecht’ [1980] Juristische Schulung 1, 9; see also RA Hillman, ‘Article 29(2) of the United Nations
Convention on Contracts for the International Sale of Goods: A new effort at clarifying the legal effect of “no
oral modification” clauses’, 21 Cornell International Law Journal 449, 451 (1988). In contrast to many others,
Hillman correctly recognizes the ambivalence of the ‘freedom of contract’ argument in the context of NOM-
clauses (eg at 449–50, 462, 463), on which the text (below 4) will elaborate. The ambivalence arises because
freedom of contract comprises both the ‘freedom of form’ and the ‘freedom of modification’ (see generally J
Basedow, ‘Freedom of Contract’ [2008] European Review of Private Law 901, 906–7)
11
 ‘Every such agreement is ended by the new one which contradicts it.’ [. . . ] What is
excluded by one act is restored by another. You may put it out by the door; it is back
through the window. Whenever two men contract, no limitation self-imposed can
destroy their power to contract again.33
Similarly, Justice Musmanno on the Pennsylvania Supreme Court wrote in Wagner v
Graziano:
The most ironclad written contract can always be cut into by the acetylene torch of
parol modification supported by adequate proof. [. . .] The hand that pens a writing
may not gag the mouths of the assenting parties. The pen may be more precise in
permanently recording what is to be done, but it may not still the tongues which
bespeak an improvement in or modification of what has been written.34
This practice of giving no effect to NOM-clauses reflects the common law position in most
of the states in the USA. 35

England & Wales

Nor is this reasoning confined to the United States. In the English case World Online v I-
Way, decided in 2002, the Court of Appeal used words remarkably similar to the ones
employed by Judge Cardozo in Beatty v Guggenheim Exploration:

33
Beatty v Guggenheim Exploration, 225 NY 380, 387–8, 122 NE 378, 381 (1919).
34
Wagner v Graziano, 390 Pa 445, 448, 136 A.2d 82, 83–4 (1957).
35
eg Alabama: Ex parte Coleman, 861 So.2d 1080, 1084–5 (Ala 2003); Indiana: City of Indianapolis v Twin
Lakes Enterprises, 568 N.E.2d 1073, 1084–5 (Ind App 1 Dist 1991); Iowa: Whalen v Connelly, 545 N.W.2d
284, 291 (Iowa 1996); Kansas: Coonrod & Walz Const Co v Motel Enterprises, 217 Kan 63, 73, 535 P.2d 971
(1975) and Saddlewood Downs, LLC v Holland Corp, 33 Kan App.2d 185, 194 (2004); Massachussetts:
Cambridgeport Savings Bank v Boersner, 413 Mass 432, 439, 597 N.E.2d 1017, 1022 (1992); Michigan:
Quality Products and Concepts v Nagel Precision, 469 Mich 362, 364, 666 N.W.2d 251, 253 (2003) (‘We
hold that parties to a contract are free to mutually waive or modify their contract notwithstanding a written
modification or anti-waiver clause because of the freedom to contract.’) (see also the discussion of Michigan
Law in VA Wellman, ‘The Unfortunate Quest for Magic in Contract Drafting’ 52 Wayne Law Review 1101,
1112–21 (2006)); Mississippi: Singing River Mall Co v Mark Fields, 599 So.2d 938, 946 (Miss 1992); Oregon:
Bennett v Farmers Ins Co of Oregon, 332 Or 138, 147, 26 P.3d 785, 792 (2001); Pennsylvania: Wagner v
Graziano, 390 Pa 445, 136 A.2d 82 (1957) and In re Franks, 95 BR 346, 352–3 (US Bankruptcy Court ED Pa
1989); Rhode Island: Fondedile v CE Maguire, 610 A.2d 87, 92 (RI 1992); Texas: Autotrol Corp v
Continental Water Sys Corp, 918 F2d 689, 692 (7th Cir 1990); Wisconsin: S & M Rotogravure Serv. v Baer,
77 Wis.2d 454, 252 N.W.2d 913, 920 (1977) and Haynes Trane Service Agency v American Standard, 51 Fed
Appx 786, 793 (10th Cir 2002). The Restatement (Second) of Contracts reflects this common law rule in
deliberate, but uncharacteristically unreasoned, departure from UCC § 2-209(2), see § 148 with comment b.
The late Professor Farnsworth explained that this innovation by the UCC was not ‘regarded suitable for a
restatement’ (emphasis added), see EA Farnsworth, ‘Ingredients in the redaction of the “Restatement (Second)
of Contracts” ’, (1981) 81 Columbia Law Review 1, 11–2. The same author points out that the common law
cases predominantly concerned cases in which there was reliance; in the absence of reliance, Farnsworth
contends that even under the common law oral modifications would not be given effect where the contract
included a NOM-clause, Farnsworth, Contracts, § 7.6 (text accompanying fn 11, citing to Foster Wheeler
Envirespouse v Franklin County Convention Facilities Authority, 678 N.E.2d 519 (Ohio 1997)).
12
 [T]he parties have made their own law by contracting, and can in principle unmake
or remake it. . . . A consensual oral variation, after all, is also an exercise of freedom
of contract.’36

Germany
In Germany, the code provision pertinent to NOM-clauses is § 125 of the German Civil
Code: ‘A juridical act that lacks the form prescribed by statute is void. In case of doubt,
non-compliance with a form requirement that is established by a juridical act likewise
results in voidness.’37 The second sentence of this provision sounds as if NOM-clauses were
enforceable under the German Code: a contractual NOM-clause is ‘a form requirement that
is established by a juridical act’, and an informal modification does not comply with this
form requirement. Hence, the modification should be void ‘in case of doubt’; and where the
NOM-clause provides that ‘a modification shall not be effective unless in writing’, there is
not even a doubt as to the intended effect of the clause.38
In addition to the wording of § 125 BGB, the legislative history also indicates that
the drafters intended NOM-clauses to be effective.39
Despite this seemingly clear position under the German Code, the German Imperial
Court reasoned along the same lines as the English and US judges described above, and
decided as early as 1910: ‘If the form required for a juridical act is established by the private
will of the parties, then the parties of the contract can abrogate the form requirement at any

36
World Online Telecom Ltd (formerly known as Localtel Ltd) v I-Way Ltd [2002] EWCA Civ 413 [10]
(Sedley, LJ). The relationship between World Online v I-Way and an earlier case, United Bank v Asif [2000]
WL 456 [15]–[18], [23] (CA Civ Div), which did give effect to the NOM-clause, is not entirely clear. See
below 3 (at nn 72–9).
37
§ 125 Bürgerliches Gesetzbuch (BGB); author’s translation and emphasis. The expression Rechtsgeschäft,
here translated as ‘juridical act’, is a term of art and used here similarly to the definition in the DCFR, II.-
1:101(2): ‘A juridical act is any statement or agreement, whether express or implied from conduct, which is
intended to have legal effect as such. It may be unilateral, bilateral or multilateral.’ For a discussion of the
relationship between the German concept of Rechtsgeschäft and the ‘juridical act’ in the DCFR see Schmidt,
‘Juridical Act’ [2010] Zeitschrift für Europäisches Privatrecht 304. The most important and frequent example
of a juridical act is a contract (cf DCFR II.-1:101(1)), and while the concept of the juridical act covers more
than just contracts, readers unfamiliar with the concept of a juridical act may want to read ‘contract’ instead of
‘juridical act’ as a first approximation.
38
The proviso ‘in case of doubt’ is meant to distinguish so-called ‘constitutive’ (ad substantiam or ad
sollemnitatem) form requirements from mere ‘declaratory’ (ad probationem) form clauses. The NOM-clauses
with which this paper is concerned will always be ‘constitutive’ form requirements. Declaratory form
requirements are those where the parties merely want to establish a duty to put modifications into writing, but
do not intend this to be a condition for the validity of the modification.
39
cf the debates in the First Commission, 4th session on 22 September 1876 and 23rd session on 21 November
1881, reprinted in Jakobs & Schubert, Beratung, Vol 2/1, 650, 654–8; for the explanatory notes to the First
Draft, see Motive zum Allgemeinen Theile, Vol 1, 181–4, reprinted in Mugdan, Die gesammten Materialien,
Vol 1, 452–4. See also Jakobs & Schubert, Beratung, Vol 2/1, 659–85.
13
time by contrarius consensus . . .’40 Ever since then, ‘simple’41 NOM-clauses have no
noticeable legal effect in Germany. Even where the parties did not think of the NOM-clause
when agreeing on a modification of the original contract (and therefore did not consciously
agree to abrogate or disapply the clause), German courts have generally held that the
informal modifying agreements are valid.42 In these cases, it was said that there was no
necessity even implicitly to agree on the abrogation or derogation from the NOM-clause; the
courts held that it sufficed that the parties agreed informally on the modification of the
contract itself.
The only concession that German courts make to the ‘simple’ NOM-clause is that
the will to modify the original contract must have been manifested ‘clearly’43 or ‘beyond
any doubt’.44 At first glance, this looks as if the NOM-clause might have an effect after all,

40
Reichsgericht, 23 November 1910, case 203/10 III [1911] Juristische Wochenschrift 94; similarly,
Reichsgericht, 5 July 1912, case III 63/12 [1912] Warneyer: Die Rechtsprechung des Reichsgerichts auf dem
Gebiete des Zivilrechts no 367; Bundesgerichtshof, 26 November 1964, case VII ZR 111/63 [1965] Neue
Juristische Wochenschrift 293 [1965] Wertpapier-Mitteilungen 175; Bundesarbeitsgericht, 10 January 1989,
case 3 AZR 460/87 [1989] Neue Juristische Wochenschrift 2149, 2150. See also Reichsgericht 21 March 1919,
case III 388/18, 95 Entscheidungen des Reichsgerichts in Zivilsachen 175–6; Bundesgerichtshof, 19
September 1966, case VIII ZR 106/64 [1967] Juristenzeitung 287; 29 November 1973, case VII ZR 205/71
[1974] Wertpapier-Mitteilungen 105; 2 March 1978, case III ZR 99/79, 71 Entscheidungen des
Bundesgerichtshofs in Zivilsachen 162, 164; 22 April 1982, case III ZR 122/80 [1982] Wertpapier-
Mitteilungen 902; Bundesarbeitsgericht 24 June 2003, case 9 AZR 302/02 [2003] Neue Juristische
Wochenschrift 3725, 3727; 28 October 1987, case 5 AZR 518/85 [1988] Neue Zeitschrift für Arbeitsrecht 425,
426; Bundesfinanzhof, 24 July 1996, case I R 115/98 [1997] Neue Juristische Wochenschrift 1327, 1328;
Oberlandesgericht Brandenburg, 18 July 2001, case 4 U 184/00 [1997] Neue Juristische Wochenschrift-
Rechtsprechungsreport 1673, 1674.
41
In contrast to ‘double’ NOM-clauses, to be discussed below 3. A ‘simple’ NOM-clause states that
modifications of the contract need to be in writing. A ‘double’ NOM-clause states that modifications of the
contract need to be in writing and that the NOM-clause itself can only be modified by writing.
42
The leading case is Bundesgerichtshof 26 November 1964, case VII ZR 111/63 [1965] Neue Juristische
Wochenschrift 293 [1965] Wertpapier-Mitteilungen 175; see also Bundesgerichtshof 29 November 1973, case
VII ZR 205/71 [1974] Wertpapier-Mitteilungen 105; Bundesgerichtshof 2 March 1978, case III ZR 99/76
[1978] Neue Juristische Wochenschrift 1585; Bundesgerichtshof 22 April 1982, case III ZR 122/80 [1982]
Wertpapier-Mitteilungen 902; Bundesarbeitsgericht 24 June 2003, case 9 AZR 302/02 [2003] Neue Juristische
Wochenschrift 3725, 3727 (obiter); but see Bundesarbeitsgericht 10 January 1989, case 3 AZR 460/87 [1989]
Neue Juristische Wochenschrift 2149, 2150 (conceding that the view that the parties must at least be aware of
the NOM-clause has its merits). Contra Bundesgerichtshof 30 September 1992, case VIII ZR 196/91 [1993]
Neue Juristische Wochenschrift 64, 65–6; Bundesgerichtshof 10 May 2007, case VII ZR 288/05 [2007] Neue
Juristische Wochenschrift 3712, 3713; Bundesfinanzhof 24 July 1996, case I R 115/95 [1997] Neue Juristische
Wochenschrift 1327, 1328; Oberlandesgericht Brandenburg 4 July 2007, case 3 U 186/06 [2008] BeckRS
09639 (all requiring that the parties must have at least implicitly manifested a will to disapply the NOM-
clause); Oberlandesgericht Königsberg 24 November 1939, case 2 U 205/39 [1940] Höchstrichterliche
Rechtsprechung no 349 (requiring an explicit abrogation of the NOM-clause).
43
Bundesgerichtshof 31 October 1984, case VIII ZR 226/83 [1985] Neue Juristische Wochenschrift 320, 322;
Bundesgerichtshof 15 May 1991, case VIII ZR 38/90 [1991] Neue Juristische Wochenschrift 1750, 1751;
Bundesgerichtshof 15 February 1995, case VIII ZR 93/94 [1995] Neue Juristische Wochenschrift 1488, 1489
(all requiring a ‘clearly expressed will’); see also Bundesgerichtshof 2 June 1976, case VIII ZR 97/74, 66
Entscheidungen des Bundesgerichtshofs in Zivilsachen 378, 381, [1976] Neue Juristische Wochenschrift 1395
(obiter).
44
Oberlandesgericht Cologne 9 November 1962, case 9 U 66/62 [1963] Der Betrieb 131; H Palm in Erman
Bürgerliches Gesetzbuch (12th edn, Cologne: Otto Schmidt 2008) § 125 para 9.
14
namely on the evidentiary standard of proof for the assertion that a modification was agreed
orally or by conduct.45 At closer inspection, this is not the case. Even if a written contract
does not contain a NOM-clause, it is rebuttably presumed that the document correctly
reflects the entire agreement.46 Any court, confronted with a party asserting that a written
agreement was subsequently modified, would require ‘clear’ evidence that proved that the
alleged modification had taken place – with or without a NOM-clause in the written
contract. 47 The presence or absence of a ‘simple’ NOM-clause changes nothing when
compared to any other written contract.
There are a few isolated German cases in which the judges argued that even a
‘simple’ NOM-clause should have some noticeable effects: The Imperial Court in one case
even considered giving full effect to ‘simple’ NOM-clauses; this consideration, however,

45
D Reinicke, ‘Bedeutung’ [1976] Der Betrieb 2289. Similarly, the Australian Federal Court stated in GEC
Marconi Systems Pty Ltd v BHP Information Technology Pty Ltd, 128 FCR 1, 62 [221] (2003): ‘Though
lacking legal effect in the face of a subsequent oral or implied agreement, it seems to be accepted that a no oral
modification clause can have significant evidentiary effect.’ (citing, inter alia, to Art 2:106 of the PECL, the
predecessor of DCFR II.-4:105). See also AW Katz, ‘The Economics of Form and Substance in Contract
Interpretation’ 104 Columbia Law Review 496, 508 (2004) (‘the presence of such a clause certainly raises the
bar of persuasion’); but Katz implicitly argues for the enforceability of NOM-clauses to allow parties ‘to
choose between relatively formal and relatively substantive interpretive regimes, and to have their choices
matter . . .’
46
‘Vermutung der Vollständigkeit und Richtigkeit’, see Bundesgerichtshof 5 July 2002, case V ZR 143/01
[2002] Neue Juristische Wochenschrift 3164–5; Bundesgerichtshof 5 February 1999, case V ZR 353/97 [1999]
Neue Juristische Wochenschrift 1702; Hertel, Staudingers Kommentar, § 125 para 92–6; Ellenberger, Palandt
§ 125 para 21; O Meyer, ‘Die privatautonome Abbedingung der vorvertraglichen Abreden –
Integrationsklauseln im internationalen Wirtschaftsverkehr –’, (2008) 72 RabelsZ 562, 584–5. Contrast the
Common Law ‘parol evidence rule’, which does not apply to subsequent conduct or statements, see below
n 107.
47
H Kötz, ‘Anmerkung zu BGH v. 19.9.1966 – VIII ZR 106/64’ [1967] Juristenzeitung 288; H Böhm, ‘Das
Abgehen von rechtsgeschäftlichen Formgeboten’ (1979) 179 Archiv für die civilistische Praxis 425, 431. Once
one allows evidence for an oral modification to be adduced, the case will usually turn on whether the trial
judge believes the witness who claims that the modification took place, or the witness who claims that the
modification did not take place (or that it had a different content). This dependence on unreliable witness
testimony is often precisely the situation the parties wanted to avoid by introducing the NOM-clause. Note that
in Germany the standard of proof is that of the judge’s ‘full conviction’ in all cases; there is no distinction into
a ‘balance of probabilities’ or ‘preponderance of evidence’ standard and a ‘clear and convincing evidence’ or
‘beyond a reasonable doubt’ standard (the only exception is that in matters of interim relief, the standard is one
of ‘plausibility’ rather than ‘conviction’).
Similarly to the German case law, common law jurisdictions will frequently require ‘clear and
convincing evidence’ where claims of an oral variation of a written contract (not containing a NOM-clause)
are made, see eg for California: Houghton v Lawton, 63 Cal App 218, 223, 218 P 475; Columbia Casualty Co
v Lewis, 14 Cal App.2d 64, 72, 57 P.2d 1010 (Ct App 1936): ‘Before courts will set aside solemn binding
written contracts and substitute therefor oral agreements, proof of the latter as to every element thereof, as well
as execution, must be clear and convincing.’; Fanucchi & Limi Farms v United Agri Products, 414 F.3d 1075,
1081–2 (9th Cir 2005).
Also cf the Italian Codice Civile, which contains in Art 2723 the following provision: ‘When it is
alleged that, after the drawing [up] of a document, a stipulation has been made, in addition or contrary to its
contents, the judge can admit proof by witnesses only if, in consideration of the character of the parties, the
nature of the contract, and any other circumstances, it appears likely that verbal additions or modifications
have occurred.’ (translation by A Monti, in: L Antoniolli & A Veneziano (eds), Principles of European
Contract Law and Italian Law (The Hague: Kluwer Law International 2005) 105–6).
15
remained obiter because in the case before the court there was anyway a lack of sufficiently
‘clear’ evidence showing ‘beyond a doubt’ that there had been a subsequent agreement on a
modification.48 Some courts have demanded that the parties not only agree to modify the
original contract in substance, but that they must additionally manifest the will to abrogate
or disapply the NOM-clause.49 An Appeals Court once required an explicit abrogation of the
NOM-clause.50 None of these exceptional cases has, however, had a lasting influence on the
German case law on ‘simple’ NOM-clauses. The case law denying effect to ‘simple’ NOM-
clauses has met with criticism, but also with approval in the legal academic literature.51
As in the US cases described above and possibly in England after World Online,52 a
‘simple’ NOM-clause in Germany will not be an effective safeguard against subsequent oral
modifications or modifications by conduct. Other countries, such as Austria, Australia, and
the Netherlands, argue along the same lines: freedom of contract allows the contractual
parties to modify their agreement informally even where there is a NOM-clause.53 The
contractual parties – and perhaps even some of their legal advisors – may be surprised to
learn that their NOM-clause was not worth the paper it was written on.54

48
Reichsgericht 23 November 1942, case V 63/42 [1943] Deutsches Recht 487, 488.
49
Bundesgerichtshof 30 September 1992, case VIII ZR 196/91 [1993] Neue Juristische Wochenschrift 64, 65–
6; Bundesgerichtshof 10 May 2007, case VII ZR 288/05 [2007] Neue Juristische Wochenschrift 3712, 3713
[26] (for the special circumstances of this case, see below n. 211); Bundesfinanzhof 24 July 1996, case I R
115/95 [1997] Neue Juristische Wochenschrift 1327, 1328; Oberlandesgericht Brandenburg 4 July 2007, case
3 U 186/06 [2008] BeckRS 09639. J Schmidt-Salzer, ‘Rechtsprobleme der Schriftformklauseln’ [1968] Neue
Juristische Wochenschrift 1257, 1258, likewise considered it essential for a valid modification that the parties
think of the NOM-clause when agreeing on the modification; however, he propagated an expansive application
of the venire contra factum proprium exception (see below 8), so that in the end he also considered ‘simple’
NOM-clauses to be ineffective (ibid at 1259).
50
Oberlandesgericht Königsberg 24 November 1939, case 2 U 205/39 [1940] Höchstrichterliche
Rechtsprechung no 349; accord W J Hau, Vertragsanpassung und Anpassungsvertrag (Tübingen: Mohr
Siebeck 2003) 58; P Mankowski ‘Anmerkung zu BAG 16 May 2000, 9 AZR 245/99’ [2001] Juristenzeitung
357, 358. But see H Kötz, ‘Anmerkung zu BGH v. 19.9.1966 – VIII ZR 106/64’ [1967] Juristenzeitung 288.
51
For criticism, see Böhm, ‘Abgehen’ (1979) 179 Archiv für die civilistische Praxis 425; H Kötz,
Vertragsrecht (Tübingen: Mohr Siebeck 2009) paras 189–96; D Einsele in Münchener Kommentar zum BGB,
Vol 1 (4th edn, Munich: CH Beck 2001) § 125 para. 66; Roquette ‘Formlose Aufhebung der
Schriftformklausel bei Mietverträgen’ (1940) 10 Deutsches Recht 961–2; F Wagner-von Papp, ‘Die
privatautonome Beschränkung der Privatautonomie’ (2005) 205 Archiv für die civilistische Praxis 342, 344–
64; for approval of the case law, see eg Palm in Erman § 125 para 9; Reinicke, ‘Bedeutung’ [1976] Der
Betrieb 2289 (both considering even double NOM-clauses to be ineffective).
52
See below 3 at nn 72–9.
53
For Australia: GEC Marconi Systems Pty Ltd v BHP Information Technology Pty Ltd, 128 FCR 1, 61–3
[213]–[223] (Federal Court of Australia 2003). This court recognized the underlying problem, described below
4; it cited to US law (Beatty) and to Art 2:106 of the Principles of European Contract Law (PECL), the
predecessor of DCFR II.-4:105, the provision discussed below 10. For Austria, see below n 243. For the
Netherlands, see WM Schrama in D Busch, EH Hondius, HJ van Kooten, HN Schelhaas & WM Schrama (eds),
The Principles of European Contract Law and Dutch Law – A Commentary (The Hague: Kluwer Law
International, 2002) 98–100.
54
cf DV Snyder, ‘The law of contract and the concept of change: Public and private attempts to regulate
modification, waiver, and estoppel’ [1999] Wisconsin Law Review 607, 639: ‘A few modern lawyers, and
certainly many of their clients, would be surprised to learn that the treatment of NOM and NOW [no-oral-
16
3. The Freedom of Contract Argument for Enforcing NOM-clauses
Some of the jurisdictions mentioned above, such as Germany and today all the states in the
US, take the non-enforceability of NOM-clauses only as the starting point for the analysis
and in some circumstances give effect to NOM-clauses. In England, at least one Court of
Appeal decision enforced a NOM-clause. International legal instruments, such as the CISG
and the UNIDROIT Principles, provide for the general enforceability of NOM-clauses.55
Just as the opponents of the enforceability of NOM-clauses did, proponents of the
enforceability of NOM-clauses invoke the principle of freedom of contract.56 They argue
that the parties have, by including the NOM-clause, contractually agreed not to give effect
to subsequent informal modifications. If subsequent informal modifications were
nevertheless enforced, the NOM-clause would have no legal effect whatsoever. The legal
system would deny the parties the enforcement of a clause agreed between the parties, even
though no public policy reason for denying the enforceability is apparent. The parties’
freedom of contract is limited in so far as they cannot agree on a legally enforceable NOM-
clause.

waiver] clauses under the common law has not changed much . . . This rule may seem odd, especially in
contract law, which devotes much of its energy to giving effect to parties’ agreements.’; see also idem,
comment in the question and answer session, (2003) 34 University of Toledo Law Review 728, 734–5: ‘These
unenforceable provisions sometimes mislead the parties. They think these clauses are enforceable. . . A lot of
us would have thought they were enforceable. . . I certainly didn’t know they weren’t enforceable at common
law. I’d managed to pass the bar, get out of law school, and everything.’; Wellman, ‘Unfortunate Quest’ 52
Wayne Law Review 1101, 1114 (2006), stating – with reference to the 125 year pedigree of the common law
rule in Michigan that NOM-clauses are ineffective –: ‘But the longevity of this view must mean that
practitioners have attempted – for that same century and a quarter – to include such clauses in their contracts.
And that, in turn, indicates that drafters have expected – for the same century and a quarter – that such clauses
would actually accomplish some useful purpose.’
55
CISG Art 29(2), UNIDROIT Principles Art 2.1.18 (which is largely identical to its predecessor provision,
Art 2.18 of the UNIDROIT Principles for International Commercial Contracts 1994).
56
eg G Schiemann ‘Das Rechtsgeschäft’ in J. von Staudingers Kommentar zum Bürgerlichen Gesetzbuch –
Eckpfeiler des Zivilrechts (Berlin: Sellier – de Gruyter, 2008) 43, 91: ‘The completely unrestricted
modifiability of agreed form requirements is not the consistent application of private autonomy, but rather the
opposite. Just as the autonomy is brought to full effectiveness by letting the decisional freedom result in the
binding effect of the decision, the self-binding effect of agreeing on a form requirement is the expression of a
responsible use of freedom [. . .].’ (author’s translation); similarly D Medicus, Allgemeiner Teil des BGB (9th
edn, Heidelberg: CF Müller, 2006) para 643; Hau, Vertragsanpassung 58; Böhm, ‘Abgehen’ (1979) 179
Archiv für die civilistische Praxis 425, 429; D Einsele in Münchener Kommentar zum BGB, Vol 1 (4th edn,
Munich: C H Beck, 2001) § 125 para 66; FW Eisner, ‘Die Schriftformklausel in der Praxis’, [1969] Neue
Juristische Wochenschrift 118, 120. See also the German decision Bundesgerichtshof 2 June 1976, case VIII
ZR 97/74, 66 Entscheidungen des Bundesgerichtshofs in Zivilsachen 378, 382, [1976] Neue Juristische
Wochenschrift 1395, described in more detail in the text below. While Hillman, ‘Article 29(2)’, 21 Cornell
International Law Journal 449 (1988), is critical towards NOM-clauses (ibid, 463), he acknowledges (at 450)
that ‘the principle of freedom of contract seems to suggest that parties should be free to include the
provision . . .’ (recognizing the ambivalence of the argument, see references in n 32). Eidenmüller, ‘Party
autonomy’ [2009] European Review of Contract Law 109, 117, cites DCFR II.-4:105 as one of the provisions
which erode party autonomy (under IV.1.), because agreed ‘form requirements are ineffective in certain cases’
(referring to DCFR II.-4:105(2)).
17
 This consideration – along with the commercial demand for NOM-clauses – has led
several jurisdictions to give effect to NOM-clauses.

Germany
As mentioned above, the courts in Germany have held since 1910 that as a general rule a
modifying oral agreement, or even conduct, can override a ‘simple’ NOM-clause.
There is, however, a commercially important exception to this general rule, an exception
whose scope appears to be growing. In 1976, the German Federal Court of Justice decided
that where:
(1) the NOM-clause provides not only that the modification of the contract, but also that
the modification of the NOM-clause itself needs to be in writing (so-called
‘qualified’ or ‘double’ NOM-clause);
(2) the parties are merchants; and
(3) the NOM-clause is individually negotiated,
such a ‘double NOM-clause’ is given full effect.57 The Court reasoned:
[The double NOM-] clause evidently has the sole purpose of preventing the
undermining of the NOM-clause that would result from binding the contractual
partners to their oral statements or even their mere conduct. There are no legal
objections against such a clause, provided it is agreed in an individually negotiated
agreement among merchants.
Freedom of contract allows them to subject their legal relationship rigidly to a
specific form. . . . If merchants . . . decide to bind themselves voluntarily in this
respect, because the associated advantage always to have certainty about the content
of contracts outweighs the disadvantage of less flexibility in the day-to-day business,
then this is to be strictly respected precisely because of the principle of freedom of
contract.58

57
Bundesgerichtshof 2 June 1976, case VIII ZR 97/74, 66 Entscheidungen des Bundesgerichtshofs in
Zivilsachen 378, [1976] Neue Juristische Wochenschrift 1395. For further decisions enforcing double-NOM
clauses see eg Bundesgerichtshof 17 September 2009, case I ZR 43/07 [2010] Multimedia und Recht 336;
Kammergericht 18 August 2005, case 8 U 106/04 [2005] BeckRS 10787, [2006] OLG Report KG 86;
Landgericht Cologne 7 January 2010, case 8 O 120/09 [2010] BeckRS 00846; see also Bundesarbeitsgericht
20 May 2008, case 9 AZR 382/07 [2009] Neue Juristische Wochenschrift 316, 317 [18] (obiter). The use and
effectiveness of double NOM-clauses had been proposed by R Boergen, ‘Die Effektivität vertraglicher
Schriftformklauseln’ [1971] Betriebs-Berater 202, 204.
58
Bundesgerichtshof 2 June 1976, case VIII ZR 97/74, 66 Entscheidungen des Bundesgerichtshofs in
Zivilsachen 378, 381–2, [1976] Neue Juristische Wochenschrift 1395 (emphasis added).
18
Many commentators welcomed the partial acknowledgement of the enforceability of NOM-
clauses, but criticized the restriction of this exception to merchants and/or to double NOM-
clauses.59
While the courts continue to require the NOM-clause to be a ‘double’ NOM clause,
in the twenty-first century there has been movement with regard to the restriction to
merchants. In 2003, the Federal Labour and Employment Court applied the exception to a
double NOM-clause between employer and employee, ie in a case in which not both parties
were merchants.60 The Court did not elaborate on the reasons for this extension, but since
2003, at least the labour and employment courts apply the double NOM-clause exception to
merchants and non-merchants alike.61
Outside the labour and employment law context, the Federal Court of Justice
decided in 2009 that double NOM-clauses are enforceable ‘at least’ where the clause is
agreed between merchants or businesses; but the dispute in question was between
businesses, so that the remark was obiter and it remains unclear whether the formulation
intended to indicate a willingness to extend the use of double NOM-clauses to non-
merchants in the future.62

United States
Many states in the US have adopted statutory law giving effect to NOM-clauses. In New
York, where the case Beatty v Guggenheim mentioned above had been decided in 1919, the
state legislature later adopted § 15-301 of the New York General Obligations Law:
§ 15-301 (1) A written agreement or other written instrument which contains a
provision to the effect that it cannot be changed orally, cannot be changed by an

59
For full effectiveness of ‘simple’ NOM-clauses: Böhm, ‘Abgehen’ (1979) 179 Archiv für die civilistische
Praxis 425; Einsele in Münchener Kommentar § 125 para 66; Roquette ‘Formlose Aufhebung’ (1940) 10
Deutsches Recht 961–2; Wagner-von Papp, ‘Privatautonome Beschränkung’ (2005) 205 Archiv für die
civilistische Praxis 342, 344–64. For an extension of the double-NOM clause case law to non-merchants:
Larenz & Wolf, Allgemeiner Teil § 27 para 85 in fn 85; Schiemann Staudinger – Eckpfeiler 91; Medicus,
Allgemeiner Teil para 643; Hau, Vertragsanpassung 58. Contra, arguing for the ineffectiveness even of double
NOM-clauses individually agreed between merchants, eg Palm in Erman § 125 para 9 (‘since the parties
cannot forgo their future freedom of contract’).
60
Bundesarbeitsgericht 24 June 2003, case 9 AZR 302/02 [2003] Neue Juristische Wochenschrift 3725, 3727.
61
eg Landesarbeitsgericht Rostock 22 April 2004, case 1 Sa 342/03; Landesarbeitsgericht Kiel 7 January 2004,
case 3 Sa 426/03.
62
Bundesgerichtshof 17 September 2009, case I ZR 43/07 [2010] Multimedia und Recht 336, 337 para 21,
referring to the abovementioned decision by the Federal Labour and Employment Court.
19
 executory agreement unless such executory agreement is in writing and signed by
the party against whom enforcement of the change is sought or by his agent.63
Several other states, such as California, have adopted similar provisions in their state laws.64
The Uniform Commercial Code provides for the effectiveness of NOM clauses in
sales of goods contracts in UCC § 2-209(2):
An agreement in a signed record which excludes modification or rescission except
by a signed record cannot be otherwise modified or rescinded, but except as between
merchants such a requirement in a form supplied by the merchant must be separately
signed by the other party.65
Statutory exceptions swallow the general rule under the common law in many cases:66 For
the sale of goods, UCC § 2-209(2) makes NOM-clauses enforceable practically across the
entire United States.67 UCC § 2A-208 establishes a parallel rule for leases, and § 303 of the
Uniform Computer Information Transactions Act (UCITA) does so for ‘computer
information transactions’.68
Outside the scope of UCC Article 2, the situation differs from state to state, but even
here statutes – such as the ones cited for two of the most populous states California and New
York –69 often provide for the enforceability of NOM-clauses. Occasionally, courts give
effect to NOM-clauses even in the absence of such statutory provisions. In Martinsville
Nylon Employees the Court of Appeals for the DC Circuit was of the view that NOM-

63
For the full text of the provision, see below n 170. For a recent application of the provision, see, eg, We’re
Associates v Hornell Brewing, 2009 WL 3066664 (NY Dist Ct 2009). See also Rose v Spa Realty Associates,
42 N.Y.2d 338, 397 N.Y.S.2d 922 (1977).
64
§ 1698(c) of the Californian Civil Code provides: ‘Unless the contract otherwise expressly provides, a
contract in writing may be modified by an oral agreement supported by new consideration.’ (emphasis added);
for applications of this provision see Marani v Jackson, 183 Cal App.3d 695, 704, 228 Cal Rptr 518 (Cal Ct
App 1986); Fanucchi & Limi Farms v United Agri Products, 414 F.3d 1075, 1081 (9th Cir 2005); Davidson v
Conocophillips, 2009 WL 2136535 (ND Cal 2009).
65
The only change affecting UCC § 2-209(2) in the 2003 revision was to substitute ‘signed record’ for
‘writing’. No change in substance was intended, see JD Wladis, ‘The contract formation sections of the
proposed revisions to UCC Article 2’ (2001) 54 SMU Law Review 997, 1037–8.
66
cf Martinsville Nylon Employees Council Corp v National Labor Relations Board, 969 F.2d 1263, 1267 (DC
Cir 1992): ‘In modern times the common law rule has been discarded in broad areas’ (Judge DH Ginsburg).
67
Only Louisiana has not adopted any version of Article 2 of the UCC.
68
Neither Article 2A of the UCC, nor the UCITA enjoys the same widespread adoption by the state
legislatures as does Article 2 of the UCC. Accordingly, I will focus on UCC § 2-209, and disregard UCC
§ 2A-208 and § 303 of the UCITA in the remainder of this paper.
69
California has a population of approximately 36 million, New York of approximately 20 million. Together,
they account for approximately 1/6 of the US population. Furthermore, even out-of-state parties often choose
New York law as the applicable law.
20
clauses are to be given effect under federal common law, at least if coupled with a merger
clause.70

England & Wales

In the English World Online case in England, while rejecting the enforceability of the
NOM-clause, Lord Justice Sedley conceded that ‘[a]mong other things, far from fettering
their freedom of contract, [counsel for the party that wants to rely on the NOM-clause] can
legitimately say that a preclusive clause like clause 21.1 [the NOM-clause] gives effect to
that freedom.’71
While World Online denied the enforceability of NOM-clauses, the Court of Appeal
had in an earlier case enforced a clause that provided that a deed could not be modified
except by a written instrument signed by the parties. In United Bank v Asif, Lord Justice
Thorpe, with whom Lord Justice Mantell agreed, said: ‘The alleged agreement constituted a
variation but was not made, as it was required to be made by clause 14.8, in writing. That
requirement was never waived by the Bank.’72 And he quoted Lord Justice Sedley as stating,
when ruling on the application to give permission to appeal: ‘Sedley LJ said, succinctly but
tellingly: “Wright J was incontestably right in concluding . . . that no oral variation of the
written terms could have any legal effect . . .” ’73
It is difficult to reconcile World Online with Asif. In Spring Finance, the judge chose
to follow World Online, because of the ‘somewhat summary approach in Asif’.74 With
respect, while the reasoning in World Online was admittedly more explicit than in Asif, the
Court of Appeal in Asif considered the NOM-clause to be ‘the real problem’ of the case;75

70
Martinsville Nylon Employees Council Corp. v National Labor Relations Board, 969 F.2d 1263, 1268 (DC
Cir 1992): ‘Whatever the ultimate merits of the common law rule denying effect to the no-oral-modifications
clause, by including the entire agreement clause the parties here made clear beyond doubt their intention not to
be bound to any informal arrangement . . . Accordingly, in this instance, the “no oral modifications” clause of
Section 22(a) and the “entire agreement” clause of Section 22(d) forestall any alteration to the Contract
through oral modifications or subsequent performance.’; quoted verbatim and followed by Judge Kollar-
Kotelly in Ghahremani v Uptown Partners, 2005 WL 3211463 (DDC 2005). Martinsville Nylon Employees
and Ghahremani thus appear to distinguish the ‘zipper clause’ (see n 109) from ‘mere’ NOM-clauses, which
were held to be unenforceable in the District of Columbia in Puma v Sullivan, 746 A.2d 871, 875 (DC 2000);
Clark v Clark, 535 A.2d 872, 876 (DC 1987); Nickel v Scott, 59 A.2d 206, 207 (DC App 1948).
71
World Online v I-Way [2002] EWCA Civ 413 [10].
72
United Bank v Asif [2000] WL 456 [23] (CA Civ Div).
73
Ibid, [17].
74
Spring Finance Limited v HS Real Company LLC [2009] EWHC 3580 (Comm), 2010 WL 902983 [36].
75
United Bank v Asif [2000] WL 456 [15]: ‘[Judge Wright’s] principal ground of rejection was that none of the
defendants’ assertions went near to meeting what he described as “the real problem”. Of course the real
problem is that the terms of the primary deed of 16th February 1998, varied only by the second deed, provide
by paragraph 14.8: “. . . No variation of this Deed shall be valid or effective unless made by one or more
instruments in writing signed by the parties to this Deed.” ’
21
and in a decision disposing of a dispute over £6 million, it is hardly conceivable that the
Court dealt ‘summarily’ with the ‘real problem’.
The Court in Spring Finance rightly rejected the attempt to distinguish Asif on the
ground that it dealt with a guarantee and should be confined to guarantees.76 Nor does it
seem possible to argue successfully that the NOM-clause in Asif was contained in a deed
and should therefore be treated differently.77 Nor can it be convincingly argued that in Asif
the additional issue, namely that the person negotiating the alleged modification may have
lacked authority, was dispositive: The Court explicitly rejected this possibility.78
Given that the Court in World Online would have been bound by the Asif precedent
and that it does not seem possible to distinguish the cases convincingly, it may be preferable
to limit World Online to the alternative ground advanced by the Court, namely that the party
invoking the NOM-clause in that case, World Online, was estopped from doing so.79 If this
were the proper reading of the cases, then NOM-clauses would be enforceable in England,
so that oral modifications could not operate as modifications, but a party relying on a
modification might be able to invoke the equitable doctrine of estoppel.
Absent any Court of Appeal or Supreme Court decision reconciling or deciding
between World Online and Asif, the position of English law on NOM-clauses remains
unsettled.

CISG and UNIDROIT Principles

When considering a European harmonization of a rule on NOM-clauses, the position of the
CISG is of particular interest. Of the 27 Member States of the European Union, 23 are
Contracting States under the CISG; the United Kingdom, the Republic of Ireland, Malta,
and Portugal are the only EU Member States that are currently not CISG Contracting
States.80 And owing to the large volume of intra-EU trade and CISG Article 1(1)b, even
persons with their place of business in one of the four non-Contracting EU Member States
may become parties to a contract to which the CISG applies.

76
Spring Finance v HS Real Company [2009] EWHC 3580 (Comm), 2010 WL 902983 [34]–[35].
77
Such a distinction would not be tenable after Berry v Berry (see above n 21).
78
United Bank v Asif [2000] WL 456 [23]: ‘The arguments as to the authority of Mr Lateef seem to me to be
of no moment.’
79
World Online v I-Way [2002] EWCA Civ 413 [13].
80
While this means that three of the four common law countries in the EU are not CISG Contracting States,
this does not indicate any dominance of Civil Law countries in the CISG context; outside the EU, many
common law countries are CISG Contracting States, such as the US, Canada, Australia, and New Zealand. The
Commission mentions only the UK, Ireland and Portugal as ‘[n]otable exceptions’ (Green Paper, above n 3, in
fn 8). Worldwide, 76 states had acceeded to the CISG at the time of writing, and in 74 of these states the CISG
had entered into force (with the Dominican Republic to follow on 1 July 2011 and Turkey on 1 August 2011).
22
The CISG takes a clear position on NOM-clauses.81 Article 29(2) provides:
A contract in writing which contains a provision requiring any modification or
termination by agreement to be in writing may not be otherwise modified or
terminated by agreement. However, a party may be precluded by his conduct from
asserting such a provision to the extent that the other party has relied on that conduct.
In pre-conference comments to the 1980 Vienna Diplomatic Conference, the United States
had stated ‘that this article on modification is of considerable practical utility’. 82 The
provision was adopted by 40 votes to 4 with 3 abstentions.83 The provision gives full effect
to NOM-clauses,84 although this is subject to an exception in the second sentence.85 In
practice, the provision does not seem to create particular problems.86

81
Generally on CISG Art 29(2) see eg Hillman, ‘Article 29(2)’ 21 Cornell International Law Journal 449
(1988); EE Bergsten, ‘Amending the Contract: Article 29 CISG’ in CB Andersen & UG Schroeter (eds),
Sharing International Commercial Law across National Boundaries: Festschrift for Albert H. Kritzer on the
Occasion of his Eightieth Birthday (London: Wildy, Simmonds & Hill Publishing, 2008) 48, 55–61, available
at <http://www.cisg.law.pace.edu/cisg/biblio/bergsten5.html>.
82
cf Secretary-General, Analysis of Comments and Proposals by Governments and International Organizations
on the Draft Convention on Contracts for the International Sale of Goods, and on Draft Provisions Concerning
Implementation, Reservations and Other Final Clauses, Document A/CONF.97/9 of 21 February 1980.
83
Summary Records of the Plenary Meetings, 7th Plenary Meeting, 8 April 1980, para 26. Sweden and
Denmark had sought to avoid being bound by what became CISG Art 29(2), see ibid, para 22.
84
P Schlechtriem tentatively suggests that ‘theoretically’ and only in exceptional circumstances, an informal
modification could circumvent the NOM-clause if interpreted as an implicit derogation from CISG Art 29(2)
under CISG Art 6, P Schlechtriem in P Schlechtriem & I Schwenzer (eds), Kommentar zum Einheitlichen UN-
Kaufrecht (4th German edn, Munich: CH Beck, 2004) Art 29 para 5; while he generally rejects such an
implicit derogation, because this ‘would be contrary to the purpose’ of the rule, he nevertheless leaves a
backdoor open by stating that ‘only the express and serious indication of such an intention should be regarded
as derogation from Article 29(2)’, see P Schlechtriem in P Schlechtriem & I Schwenzer (eds), Commentary on
the UN Convention on the International Sale of Goods (2nd English edn, Munich: CH Beck 2005) Art 29 para
5; similarly idem, in CB Andersen & UG Schroeter (eds), Sharing International Commercial Law across
National Boundaries: Festschrift for Albert H. Kritzer on the Occasion of his Eightieth Birthday (London:
Wildy, Simmonds & Hill Publishing, 2008) 416, 423 (‘it must be established that the parties indeed
consciously intended to deviate from Article 29(2) CISG’). In my view, this is still contrary to the wording and
purpose of Art 29(2), first sentence, which requires a writing, and not only an ‘express and serious indication’;
cf A Geldsetzer, Einvernehmliche Änderung und Aufhebung von Verträgen (Baden-Baden: Nomos, 1993) 151;
W Wasmer, Vertragsfreiheit im UN-Kaufrecht (Hamburg: Kovac, 2004) 85–6. Everything short of an
agreement in writing may at best qualify as ‘conduct’ on which the other party may rely in certain
circumstances, see Art 29(2), second sentence, and below 8. P Schlechtriem & P Butler, UN Law on
International Sales (Berlin: Springer, 2009) para 99 are not entirely clear on this point.
85
For the exception in the second sentence of CISG Art 29(2) see below 8.
86
An Arbitral Tribunal at the ICC International Court of Arbitration in Zurich remarked that the merger clause
and the NOM-clause are ‘typical clauses, and there can be no doubt for any party engaged in international
trade that the clauses mean, and must mean, what they say.’ Arbitral Award no 9117, March 1998,
<http://cisgw3.law.pace.edu/cases/989117i1.html>, reprinted in MJ Bonell, The UNIDROIT Principles in
Practice (2nd edn, Ardsley: Transnational Publishers, 2006) 727. In Graves Import v Chilewich International,
1994 WL 519996 (SDNY 1994), <http://cisgw3.law.pace.edu/cases/940922u1.html>, the Court declined the
invitation to go into the merits of an assertion that the contract had been modified orally by referring to Art
29(2) of the CISG (text at and in fn 2). In an Arbitral Award of the ICC International Court of Arbitration
(Award no 11849 at [31]–[32], <http://www.unilex.info>), the arbitrator enforced the NOM-clause (apparently
referring exclusively to Art 2.18 of the UNIDROIT Principles 1994 – the predecessor of Art 2.1.18
UNIDROIT Principles 2004 –, rather than to Art 29(2) of the CISG, despite the fact that the CISG was
applicable to the contract). Other CISG cases enforcing NOM-clauses include China International Economic &
23
Similarly, the UNIDROIT Principles provide in Article 2.1.18:
A contract in writing which contains a clause requiring any modification or
termination by agreement to be in a particular form may not be otherwise modified
or terminated. However, a party may be precluded by its conduct from asserting such
a clause to the extent that the other party has reasonably acted in reliance on that
conduct.
Apart from the jurisdictions and legal instruments mentioned above, further jurisdictions
appear willing to enforce NOM-clauses as well.87

4. Policy Reasons for and Against the Enforceability of NOM-clauses

Freedom of Contract?
Both those supporting the enforceability of NOM-clauses and those arguing against it
invoke the principle of freedom of contract.88 At first glance, this seems to be contradictory.
At closer inspection, the contradiction resolves itself: the opposing views focus on the
freedom of contract at different points in time. Those arguing in favour of the enforceability
of NOM-clauses focus on the freedom of contract at the time of the drafting the original
contract – let us call this point in time t1. If we do not enforce NOM-clauses, we deny the
parties in t1 the option of agreeing on an effective NOM-clause. In contrast, those arguing
against the enforceability of NOM-clauses focus on the freedom of contract at the time of
the modifying agreement – let us call this point in time t2. Enforcing the NOM-clause means

Trade Arbitration Commission 16 December 1997 (hot-dipped galvanized steel coils case),
<http://cisgw3.law.pace.edu/cases/971216c1.html> (Part 4); Tribunal of International Commercial Arbitration
at the Russian Federation Chamber of Commerce and Industry 16 February 2004, Arbitration proceeding
107/2002, <http://cisgw3.law.pace.edu/cases/040216r1.html> (Part 3.4.1); Tribunal of International
Commercial Arbitration at the Ukraine Chamber of Commerce and Trade 25 November 2002,
<http://cisgw3.law.pace.edu/cases/021125u5.html> [5]. The decision Oberlandesgericht Innsbruck 18
December 2007, case 1 R 273/07t, <http://cisgw3.law.pace.edu/cases/071218a3.html> (Steel bars case)
applies CISG Art 29(2) (without finally deciding whether the CISG was applicable), but considers the form
requirement agreed between the parties to be satisfied by a written confirmation letter to which the other party
did not object (Part 2 of the judgment).
87
See below text in and at nn 237–239 for Italian law, and in and at n 231 for Scots law; also cf Art 159(1), (2)
of the Russian Civil Code (trans. in Butler, Russian Civil Code, 72): ‘(1) A transaction for which the written . . .
has not been established by law or by agreement of the parties may be concluded orally. (2) Unless otherwise
established by agreement of the parties, all transactions to be performed by those who concluded them
themselves may be concluded orally . . .’ Pursuant to Art 162 of the Russian Civil Code, the failure to comply
with such ‘written form of a transaction shall deprive the parties of the right in the event of a dispute to refer in
confirmation of the transaction and its conditions to witness testimony, but shall not deprive them of the right
also to cite written and other evidence.’
88
See above nn 32, 56 and the quotes from the cases cited in the text above 2 and 3.
24
reducing the degree of freedom in t2, because the parties are no longer free to modify the
contract without conforming to whatever form was agreed in the NOM-clause.
Both proponents and opponents of the enforceability are therefore right in invoking
the freedom of contract for their position: Enforcing the NOM-clause reduces freedom in t2,
not enforcing the clause reduces freedom in t1. The apparent paradox results from the nature
of the NOM-clause: it is intended (in t1) to reduce the freedom of the parties in the future
(t2). The NOM-clause is therefore an example for the ‘paradox of freedom’ – unlimited
freedom in t1 would seem to imply the freedom to limit one’s future freedom in t2.89
Granting such unlimited freedom in t1 is distinctly unattractive: As John Stuart Mill noted,
this would ultimately mean that a human being ‘would be free not to be free’ and could sell
himself into slavery.90
Because there is a trade-off between freedom in t1 and freedom in t2, we cannot
maximize freedom in both t1 and t2. We have to choose between reducing freedom in t1 and
reducing freedom in t2. This choice cannot be made in the abstract; in some situations, the
89
Other examples of such self-limiting agreements include:
(i) framework agreements, seeking to limit the freedom in subordinate contracts;
(ii) agreements never to divorce: see eg German Bundesgerichtshof 9 April 1986, case IVb ZR
32/85, 97 Entscheidungen des Bundesgerichtshofs in Zivilsachen 304;
(iii) self-exclusion agreements from casinos by pathological gamblers: England: Calvert v
William Hill [2008] EWCA Civ 1427, [2009] Ch 330; US: eg Merrill v Trump Indiana, 320 F.3d
729 (7th Cir 2003); Stulajter v Harrah’s Indiana Corp., 808 N.E.2d 746 (Ind App 2004);
Germany: Bundesgerichtshof, 15 December 2005, joint cases III ZR 65/05 and III ZR 66/05, 125
Entscheidungen des Bundesgerichtshofs in Zivilsachen 276, [2006] Juristenzeitung 468,
overruling Bundesgerichtshof 31 October 1995, case XI ZR 6/95, 131 Entscheidungen des
Bundesgerichtshofs in Zivilsachen 136, 1996 Neue Juristische Wochenschrift 248; G Schulze,
‘Verträge zum Schutz gegen sich selbst’, in HP Mansel et al (eds) Festschrift für Erik Jayme, Vol
II, 1577 (Munich: Sellier European Law Publishers, 2004); idem, Die Naturalobligation
(Tübingen: Mohr Siebeck, 2008) 333 et seq; see also the excellent comparative overview by S
Sendmeyer, MJ Doris, S Marco Colino, I Kull, T Katsas, E Ferrante, C Kindler, S Navas Navarro
& N Hoekx, ‘Selbstsperre beim Glückspiel in Deutschland’ [2009] European Review of Private
Law 377–451; F Wagner-von Papp, ‘Should we be free not to be free to ruin ourselves?’ in M
Freeman & O Goodenough (eds) Law, Mind and Brain (Farnham: Ashgate 2009) 81 (draft
available at <http://ssrn.com/abstract=1590483>); idem, ‘Privatautonome Beschränkung’ (2005)
205 Archiv für die civilistische Praxis 343, 364–85;
(iv) immutable contracts: contrast C Jolls, ‘Contracts as Bilateral Commitments: A new
perspective on contract modification’ (1997) 26 Journal of Legal Studies 203 with KE Davis,
‘The Demand for Immutable Contracts: Another look at the law and economics of contract
modifications’ (2006) 81 NYUL Rev 487; see also A Schwartz & J Watson, ‘The Law and
Economics of Costly Contracting’ (2004) 20 Journal of Law, Economics and Organization 2.
See generally SA Smith, ‘Future Freedom and Freedom of Contract’ (1996) 59 Modern Law Review 167; P
Saprai, ‘The Principle Against Self-Enslavement in Contract Law’ (2009) 26 Journal of Contract Law 25; K
Eggert, ‘Lashed to the Mast and Crying for Help: How Self-Limitation of Autonomy Can Protect Elders from
Predatory Lending’ (2003) 36 Loyola of Los Angeles Law Review 693; J Elster, Ulysses and the Sirens (rev
edn, Cambridge: Cambridge University Press, 1984); J Elster, Ulysses Unbound (Cambridge: Cambridge
University Press, 2000); B Kuklin, ‘Self-Paternalism in the Marketplace’ (1992) 60 University of Cincinnati
Law Review 649; TC Schelling, ‘Enforcing Rules on Oneself’ (1985) 1 Journal of Law, Economics &
Organization 357; TC Schelling, Strategies of Commitment and Other Essays (Cambridge, Mass: Harvard
University Press, 2006).
90
JS Mill, On Liberty (ed CV Shields, Upper Saddle River, NJ: Prentice Hall, 1859/1997) 125.
25
reduction of freedom in t1 may be more acceptable than a reduction of freedom in t2, in
others the opposite may be true. The decision depends on the relative degree of reduction in
freedom at the two points in time. In John Stuart Mill’s example of self-enslavement,
absolute freedom in t1 would mean extinguishing freedom of contract in t2 completely. The
law does not enforce the self-enslavement agreement and thus denies the freedom in t1 for
the sake of retaining freedom in t2. On the other hand, all contracts constrain the parties’
future freedom to a certain degree, because the promises made in t1 have to be performed in
t2 lest the party risk incurring the sanctions for a breach of contract. By enforcing contracts,
the law allows the parties in t1 to limit the parties’ freedom in t2.91
It would be wrong, however, to conclude that because all contracts constrain the
parties’ future freedom and the law enforces this self-limitation, NOM-clauses are just
another application of this principle and must likewise be enforced. The difference is that in
the case of a typical contractual promise to do or not to do something, each party’s future
freedom is reduced in that they must not unilaterally deviate from the contract; but jointly
the contractual parties can at any time decide that the promise need not be performed.92 In
contrast, in the NOM-clause both parties agree in t1 that they cannot jointly agree in t2 to
deviate from the NOM-clause, unless they conform to the agreed form. Therefore, the self-
limiting effect of the NOM-clause is greater than that of most other contractual clauses. And
generally the law has a tendency to favour the later agreement over a previous one.93
On the other hand, it should be obvious that the self-limiting effect of the NOM-
clause is considerably smaller than that of self-enslavement: all the parties need to do in t2 in
order to lift the self-limitation that was imposed in t1 is to put their modification into writing
(or whatever form was agreed). The fact that the NOM-clause only trivially reduces
freedom in t2, while leaving the parties the freedom to agree on a NOM-clause in t1, is an
argument for the enforcement of NOM-clauses. Furthermore, the argument that the law
generally favours later agreements over previous ones is true ‘[a]ll other things being equal
(or equally unknown)’;94 yet, all things are not equal if the parties specifically anticipate the
possibility of this later change of minds and explicitly provide for this contingency by

91
cf Smith, ‘Future Freedom’ (1996) 59 Modern Law Review 167–87.
92
Again: always provided that the other requirements for a binding contract are met, in particular that the
promise not to enforce the contractual promise is supported by consideration (where required).
93
Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 463 (1988): ‘All other things being
equal (or equally unknown), perhaps the law should respect the latest bargain or position of the parties’; GEC
Marconi Systems Pty Ltd v BHP Information Technology Pty Ltd, 128 FCR 1, 62 [220] (Federal Court of
Australia 2003).
94
See n 93.
26
agreeing to a NOM-clause.95 In the case of NOM-clauses, the parties have (i) anticipated the
possibility of an oral modification, they have (ii) specifically provided for this case that the
oral modification should not be binding, and they have (iii) the possibility autonomously to
reverse this self-limitation by agreeing in writing either to modify the contract – leaving the
NOM-clause in place – or to abrogate the NOM-clause.96
I believe that this is an argument for preferring the slight reduction in freedom in t2,
which lies in forcing the parties to put their agreement into writing, over the alternative of
depriving the parties of the freedom in t1 to agree on an effective NOM-clause; it is not,
however, a conclusive argument. What we can say is that asserting the ‘freedom of contract’
argument cannot be the clinching argument for either side; there is no ‘logical’ superiority
of one solution over the other. 97
Courts, and certainly legislators, are therefore free to opt for either solution without
infringing the principle of freedom of contract. The choice between the solutions has to
depend on other factors. While this insight might seem trivial, several courts – and
apparently the drafters of the DCFR98 – appear to have felt compelled by reasons of logic
not to enforce NOM-clauses.99 Judge Posner of the Seventh Circuit found characteristically
clear words on the kind of ‘reasoning’ employed by cases such as Beatty and Wagner: ‘This
is not reasoning; it is a conclusion disguised as a metaphor.’100

95
Enforcing NOM-clauses gives the parties the opportunity of including effective NOM-clauses, but does not
compel them to do so. Cf the quote of Judge Ginsburg in the text accompanying n 124.
96
I have argued elsewhere for the enforcement of self-limiting agreements that anticipate a specific change in
circumstances and deal with this contingency in advance, provided that there is a possibility of autonomous
reversal of the self-limitation. Wagner-von Papp, ‘Self-Exclusion Agreements’, in Freeman & Goodenough,
Law, Mind and Brain 81–126 (draft at <http://ssrn.com/abstract=1590483>). At the time I wrote that paper, I
had not seen Martinsville Nylon Employees Council v NLRB, 969 F.2d 1263, 1269 (DC Cir 1992), which uses
very similar language: ‘The additional requirement . . . that any further agreements be in writing indicates that
the parties anticipated the possibility of entering into further agreements and specifically provided a procedure
for incorporating them . . .’
97
I understand Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 463 (1988) to acknowledge
the inconclusive nature of the freedom of contract argument, even though his preference would be for the
enforcement of the later agreement. See also Snyder, ‘Concept of change’ [1999] Wisconsin Law Review 607,
640: ‘The question for the court is not whether to honor the parties’ original agreement, but rather which of
their agreements should be effective.’
98
Below 10.
99
See much of the case law cited above 2. For exceptions, see the English case World Online [2002] EWCA
Civ 413 and the Australian case GEC Marconi Systems Pty Ltd v BHP Information Technology Pty Ltd, 128
FCR 1, 62 [220] (Federal Court of Australia 2003) (citing the passage from Snyder’s article quoted above in
n 97).
100
Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280, 1286 (7th Cir 1986) (Judge Richard
Posner).
27
Commercial Need
A first indication for the commercial demand for NOM-clauses is the – admittedly
unsystematic – empirical evidence that they are frequently used in business relationships,101
even in jurisdictions in which their enforcement is doubtful.102
The reasons for which parties employ NOM-clauses vary. Using Lon Fuller’s
framework, one can distinguish evidentiary, cautionary and channelling functions of form
requirements,103 and the parties may want to use NOM-clauses for any or all of these
purposes.
Turning first to the evidentiary function, the parties may want to be certain that they
have a written version of the entire agreement at all times, especially in long-term
contractual relationships, or where a party enters into a multitude of similar transactions.104
This is especially important in business relationships where the agents for the parties may
change over time; it may be difficult or impossible to prove or refute the existence or
content of a modification if the agents are no longer in the employ of the parties, unless the
modifying agreement has been reduced to writing.
In order to ensure that the parties can rely on the documents ‘on file’ to reflect the
entire agreement, in particular in Common Law jurisdictions they typically include a clause
in the contract that is – unsurprisingly – called an ‘entire agreement’ clause (or, slightly less

101
The comments to DCFR note that NOM-clauses ‘often occur, especially in long-term contracts’, C von Bar,
E Clive, H Schulte-Nölke et al, DCFR Full Edition, Vol I, 288. An Arbitral Tribunal at the ICC International
Court of Arbitration in Zurich (n 86) called the NOM-clause (‘written modification clause’) and the merger
clause ‘typical clauses’.
102
In England, where the enforcement of NOM-clauses is unclear (see above 3 at nn 72–9), it is easy to find
contracts choosing English law as the applicable law and including a NOM-clause. Beside the clauses in
World Online, Asif and Spring Finance, see eg clause 9(b) in the Master Agreement in AS Klaveness
Chartering v Pioneer Freight Futures [2009] EWHC 3386 (Comm), 2009 WL 4872647 [9]; clause 15 of the
contract in Watersheds Ltd v DaCosta [2009] EWHC 1299 (QB), [2010] Bus LR 1 [8]; clause 8 of the contract
in Resteel Trading v The Commissioners for Her Majesty’s Revenue and Customs [2009] UKFTT 236 (TC) [5];
clause 12.3 of the contract in Launahurst Limited v Mr ND Larner, Employment Appeal Tribunal, 2009 WL
2392313 [20]; the NOM-clause in the letter in Steven Andrew Clark v Nomura International [2000] WL
1213073 (QBD) [7] and [10.2]. Numerous further examples can be found in online standard terms and
conditions. P Garry & P Ashford, ‘Partnerships: Members Only’, The Lawyer 8 November 2004, 25, suggest
that because ‘[t]here is no requirement for an LLP agreement to be in writing, no bar on implied terms, and no
bar on oral or implied variation’ and ‘ordinary rules of contract apply’, ‘[e]ntire agreement clauses and
provisions limiting the manner in which variations can be effected are accordingly very important’ (emphasis
added).
103
LL Fuller, ‘Consideration and Form’, (1941) 41 Columbia Law Review 799, 800–1 and passim. One can
identify many more functions of form requirements in general (see P Mankowski, ‘Formzwecke’ [2010]
Juristenzeitung 662–668, distinguishing between 14 different functions), but for the purposes of this article,
the three ‘classic Fuller functions’ will suffice; some of the additional functions mentioned by Mankowski can
only be achieved by statutory form requirements, others may be ancillary to the three classic functions.
104
cf Bundesgerichtshof 2 June 1976, case VIII ZR 97/74, 66 Entscheidungen des Bundesgerichtshofs in
Zivilsachen 378, 382.
28
intuitively, a ‘merger’ or ‘integration’ clause). 105 Among other things, 106 this clause
facilitates invoking the parol evidence rule: The parties cannot adduce extraneous evidence
to vary, add to or contradict the terms of a completely integrated agreement, and the entire
agreement clause indicates that the agreement is a completely integrated one. One of the
many weaknesses of the parol evidence rule is, however, that it excludes only evidence
aiming to prove a supplementary or contradictory agreement that was made prior to or
contemporaneously with the written contract; it does not exclude any evidence for
subsequent modifications.107 It is this gap in the parol evidence rule that the parties seek to
close by adding a NOM-clause to the entire agreement clause. 108 The result of the
combination of these clauses is intended to be that the written document(s) can be trusted to
reflect at all times the entire agreement between the parties.109

105
cf UCC § 2-202; Art 2.1.17 of the UNIDROIT Principles; DCFR II.-4:104; C Mitchell, ‘Enitre Agreement
Clauses: Contracting Out of Contextualism’ (2006) 22 Journal of Contract Law 222; E Peden & JW Carter,
‘Entire Agreement’ (2006) 22 Journal of Contract Law 1. For an elaborate comparative account of
merger/entire agreement clauses, and in particular the differences between US and English practice, see Meyer,
‘Privatautonome Abbedingung’ (2008) 72 RabelsZ 562. A typical entire agreement clause may state: ‘The
Contract contains the whole agreement between the parties in respect of the subject matter and supersedes and
replaces all prior proposals, agreements, representations and negotiations relating thereto, whether written or
oral or implied between the parties or their respective advisers, unless otherwise provided herein.’
106
Meyer, ‘Privatautonome Abbedingung’ (2008) 72 RabelsZ 562, 582–3, rightly points out that in England
the attempt at excluding liability for representations and collateral warranties is one of the main functions of
entire agreement clauses. As AG Guest in Chitty on Contracts § 12-104 points out, the clause may be more
effective against assertions of collateral warranties than against the assertions of representations, esp
misrepresentations.
107
See the English case Goss v Lord Nugent (1833) 5 Barnewall and Adolphus 58, 64–5, 110 ER 713, 717
(KB), in which Denman, CJ, stated: ‘By the general rules of the common law, if there be a contract which has
been reduced into writing, verbal evidence is not allowed to be given of what passed between the parties,
either before the written instrument was made, or during the time that it was in a state of preparation, so as to
add to or subtract from, or in any manner to vary or qualify the written contract; but after the agreement has
been reduced into writing, it is competent to the parties, at any time before breach of it, by a new contract not
in writing, either altogether to waive, dissolve, or annul the former agreements, or in any manner to add to, or
subtract from, or vary or qualify the terms of it, and thus to make a new contract; which is to be proved, partly
by the written agreement, and partly by the subsequent verbal terms engrafted upon what will be thus left of
the written agreement.’; see also AG Guest in Chitty on Contracts § 12-111. The same general rules apply in
the US, see eg Marani v Jackson, 183 Cal App.3d 695, 699 in fn 2, 228 Cal Rptr 518 (Cal Ct App 1986);
Whalen v Connelly, 545 N.W.2d 284, 291 (Iowa 1996).
108
ICC International Court of Arbitration in Zurich (n 86): NOM clause ‘has the same effects as the merger
clause with regard to any future negotiations, promises and any other extrinsic evidence . . .’; Meyer,
‘Privatautonome Abbedingung’ (2008) 72 RabelsZ 562, 564 (NOM clauses are the ‘complement in relation to
time’ to entire agreement clauses). For an array of problems resulting from the dichotomy ‘before’ (entire
agreement clause) and ‘after’ (NOM-clause), see GS Crespi, ‘Clarifying the boundary between the parol
evidence rule and the rules governing subsequent oral modifications’, (2008) 34 Ohio Northern University
Law Review 71.
109
So frequent is the combination of the entire agreement clause with a NOM-clause that the terminology
‘zipper clause’ has developed to denote the combined clause (see Pace v Honolulu Disposal Services, 227 F.3d
1150, 1159 (9th Cir. 2000); Farnsworth, Contracts, § 7.6 fn 5). As mentioned above n 70, DC courts have
enforced such zipper clauses without finally deciding on whether isolated NOM-clauses would be enforceable.
Of course, even where the parties use a zipper clause, various other exceptions to the parol evidence rule
remain. Parties may, for example, retain the power to show by adducing extraneous evidence that the contract
is void, voidable, or not supported by consideration, how an ambiguous term is to be interpreted, and, at least
29
 Being able to rely on the evidentiary function of the documents ‘on file’ has three
advantages. First, where the parties employ agents, the parties can be certain that the agents
have not entered into any binding oral agreements without the principal’s knowledge.
Secondly, comprehensive written documentation helps to avoid disputes from
arising that might otherwise result from differences about the existence or content of
(alleged) oral agreements. If agreements have to be in writing in order to be binding, the
parties are more likely to resolve their conflicts out of court by consulting the respective
documents.
Thirdly, where a dispute does arise and reaches the courts or arbitral tribunals, the
dispute can be resolved – largely – by referring to documentary evidence. The drawbacks of
relying on witness testimony of oral negotiations are well explained in the English case
Chartbrook v Persimmon, where Lord Hoffmann stated – although in a different context,
namely the exclusion of prior negotiations as background information in contractual
interpretation – that the admission of evidence about oral negotiations:

. . . would create greater uncertainty of outcome in disputes . . . and add to the cost of
advice, litigation or arbitration. . . . [S]tatements would have to be taken from those
who took part in oral negotiations. Not only would this be time-consuming and
expensive but the scope for disagreement . . . would be considerably increased. . . .
The document should so far as possible speak for itself. As Popham C.J. said in the
Countess of Rutland’s Case (1604) 5 Co Rep 25b, 26a:
‘it would be inconvenient, that matters in writing made by advice and on
consideration, and which finally import the certain truth of the agreement of
the parties should be controlled by averment of the parties to be proved by
the uncertain testimony of slippery memory.’
. . . [S]tatements in the course of . . . negotiations will be drenched in subjectivity
and may, if oral, be very much in dispute. It is often not easy to distinguish between
those statements which (if they were made at all) merely reflect the aspirations of
one or other of the parties and those which embody at least a provisional
consensus . . .110

in California, that a term is ambiguous in the first place (Pacific Gas & Electric v GW Thomas Drayage &
Rigging Co, 69 Cal.2d 33, 442 P.2d 641 (1968); contrast the situation in New York, WWW Associates v
Giancontieri, 77 N.Y.2d 157, 566 N.E.2d 639 (1990)).
110
Chartbrook v Persimmon [2009] UKHL 38, [2009] 1 AC 1101 [36]–[38].
30
Even though these observations were made in a different context, they encapsulate many of
the reasons for which parties may want to exclude evidence on alleged oral modifications,
and why NOM-clauses make commercial sense. As long as evidence on alleged oral
modifications is admissible, there are at least three sources for conflict: (i) fraudulent
assertions of oral modifications; (ii) good-faith, but erroneous assertions of oral
modifications based on a misunderstanding of the ‘statements . . . drenched in subjectivity’;
and (iii) faulty recollection due to the witnesses’ ‘slippery memory’. These sources for
conflict will be addressed in turn.
Parties may want to establish their ‘private Statute of Frauds’, as NOM-clauses are
sometimes called,111 for the same reasons for which Parliament enacted the original Statute
of Frauds 1677. The mischief that the Statute of Frauds 1677 wanted to address was the
‘many fraudulent Practices which are commonly endeavoured to be upheld by Perjury and
Subornation of Perjury’, 112 or, in the words of Lord Bingham in Actionstrength v
International Glass Engineering, ‘the calling of perjured evidence to prove spurious
agreements said to have been made orally.’113 Mutatis mutandis, contractual parties today
may agree on a NOM-clause in order to prevent the other party from fraudulently asserting
modifications that did not take place.114 Accordingly, Judge Richard Posner, for the majority
in Wisconsin Knife, opined that ‘[t]he main purpose of forbidding oral modifications is to
prevent the promisor from fabricating a modification that will let him escape his obligations
under the contract’;115 and Judge Easterbrook in his dissent likewise noted that the NOM-
clause’s ‘principal function is to make it easier for businesses to protect their agreement
against casual subsequent remarks and manufactured assertions of alteration.’116
Even if the parties do not suspect the respectively other party of behaving in bad
faith in the future, they may fear that witnesses, who are called upon to give testimony about
an alleged oral modification in potential future litigation or arbitration, may misinterpret
mere negotiations as binding modifications. Here, the evidentiary function meets the

111
Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 450; idem, ‘How to create a
commercial calamity’, (2007) Ohio State Law Journal 335, 337; del Pilar Perales Viscasillas, ‘Modification’,
25 Journal of Law and Commerce 167, 176 (2005); see also comment 3 to UCC § 2-209 (parties may wish to
make their ‘own Statute of Frauds’). See also Farnsworth, Contracts, § 7.6, pointing out that the NOM-clause
is usually even stricter than the Statute of Frauds, as it requires the modification to be in writing, and not only
an agreement to be memorialized in writing.
112
See the preamble to An Act for prevention of Frauds and Perjuryes 1677 (29 Cha II, c 3); the short title
Statute of Frauds 1677 was introduced by the Short Titles Act 1896 c 14.
113
Actionstrength Limited v International Glass Engineering [2003] UKHL 17, [2003] 2 AC 541, 544 [1].
114
See eg Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 450 (1988).
115
Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280, 1287 (7th Cir 1986).
116
Ibid, at 1292.
31
channelling function of the form requirement: ‘It serves . . . to mark or signalize the
enforceable promise; it furnishes a simple and external test of enforceability’.117
There is a danger that parties negotiating a modification differ in their view on
whether an agreement on a modification has been reached or what the content of such a
modification is. Misunderstandings may arise, for example, from certain cognitive defects,
such as the ‘confirmation bias’: human beings tend to search for, interpret and recall
ambiguous evidence selectively, depending on whether it is in accordance with their
existing views or wishes.118 Such wishful thinking may lead the party in whose favour a
contract modification would be to assume – in good faith, but erroneously – that the
negotiations about a contract modification have resulted in an agreement, where in reality
the parties have not reached a consensus. This danger may be particularly great where the
other party was indecisive and equivocated, or where the other party framed the rejection of
a proposal in a particularly polite way that was open to misinterpretation.119 Such problems
may be exacerbated where the contracting parties come from different cultures – what is
considered to be a polite rejection of a proposal in one culture may be misunderstood as an
acceptance by a person from another culture not used to the same subtleties. A NOM-clause,
requiring the parties to reduce agreements to writing, would bring to light any such
misunderstandings.120
Even where the parties recognized at the time of the oral negotiations that a
consensus had not yet been reached, the passage of time may cloud the witnesses’
recollection. Confirmation bias may again be a contributing factor: It affects not only the
selective collection and interpretation of information, but also the selective recall of
information.121 Once again, written evidence, compelled by a NOM-clause, would ‘import
the certain truth of the agreement’.122

117
Fuller, ‘Consideration and Form’ (1941) 41 Columbia Law Review 799, 800.
118
See, eg, ME Oswald & S Grosjean, ‘Confirmation Bias’ in RF Pohl (ed), Cognitive Illusions: A handbook
on fallacies and biases in thinking, judgement and memory (Hove: Psychology Press, 2004) 79.
119
Such as the ‘subsequent casual remarks’, the protection against which is one of the principal functions
according to Judge Easterbrook (above n 116).
120
cf R von Jhering, Geist des römischen Rechts auf den verschiedenen Stufen seiner Entwicklung Part II.2
(Leipzig: Breitkopf und Härtel 1875) 495, arguing generally that form requirements allow an easy distinction
between those statements that were meant to be legally binding from those that were not meant to be binding,
eg because they were only expressing the contemplation of an agreement. Jhering gives the example of one
party saying ‘I want to sell to you’, which can either express the present intention to sell, or the contemplation
of a future sale.
121
Oswald & Grosjean, ‘Confirmation Bias’ at 87–94.
122
See above text accompanying n 110.
32
 All this is not to imply that the trend in modern legal systems to allow informal
contracts should not be welcomed.123 Where the parties to a contract, however, agree that
they prefer to recognize only written agreements as binding between themselves, so as to
minimize the possibility of fraudulent assertions, misunderstandings or faulty recollection,
there is no reason why the law should not support this commercial need by enforcing NOM-
clauses. As Judge Douglas Ginsburg of the Court of Appeals for the D.C. Circuit Court
noted in Martinsville Nylon Employees:
The drawback of the common law rule is that it denies to all contracting parties, no
matter how sophisticated, the ability to decide for themselves whether to restrict the
manner in which their agreement may be modified. The UCC rule, on the other hand,
enables those parties who mutually value certainty in their relations to have it; under
the UCC rule, moreover, no one has to agree to rule out oral modifications nor,
having so agreed, are the parties precluded from executing a mutually agreeable
modification. . . Judicial respect for a no-oral-modification clause enables parties
who value certainty to provide for it in their written agreement while taking nothing
away from parties who prefer some uncertainty, perhaps as the price of the arguably
greater flexibility attendant to oral modifications.124
At the same time, the NOM-clause has positive effects on the procedural efficiency of the
courts and tribunals. Lord Hoffmann’s exhortations in Chartbrook about the ‘cost of advice,
litigation or arbitration’ and the time-consuming nature of taking evidence when admitting
evidence about oral negotiations are apposite. Where NOM-clauses are enforced, many
conflicts can be resolved exclusively or predominantly on the basis of written evidence; and
many legal systems provide for summary procedures in such cases. Lord Justice Schiemann
acknowledged the practical virtues of a NOM-clause by stating in an obiter dictum in World
Online that he had been impressed by the submission:
that the purpose of a clause such as clause 21 [scil.: the NOM-clause] is not to
prevent the recognition of oral variations, but rather, casual and unfounded

123
See nn 17, 19.
124
Martinsville Nylon Employees Council Corp v National Labor Relations Board, 969 F.2d 1263, 1267–8
(DC Cir 1992). Also cf the remarkably similar language used by the German Bundesgerichtshof (n 58). The
view denying effect to NOM-clauses (above 2.) essentially makes non-enforceability a mandatory rule (O
Ben-Shahar, ‘Formalism in Commercial Law’ 66 University of Chicago Law Review 781, 792 (1999)) without
any good reason. From an economic perspective, the choice of an inefficient default rule would not matter in
the absence of transaction costs, because the parties can easily contract around it (‘Coase Theorem’, see R
Coase, ‘The Problem of Social Cost’ (1960) 3 Journal of Law and Economics 1); but mandatory rules prevent
such direct contracting around the legal rule. It is true that there are ways to contract around even the
mandatory ‘no effective NOM-clause rule’ (eg by interposing an agent with limited authority, below n 133),
but such contracting around is costly.
33
 allegations of such variations being made. [I]f in cases such as the present we allow
something going to trial, precisely that is allowed against which the parties may be
regarded as having sought to safeguard themselves.125

Cautionary Function?
There may be further reasons for NOM-clauses: For example, a person that is aware of her
own tendency to accept offers rashly may want to protect herself from such rash agreements
to modify a contract by inserting a NOM-clause into the contract. In this example, the
NOM-clause serves a cautionary function.126 In the majority of commercial cases, such a
cautionary function is presumably a subsidiary consideration at best. Furthermore, as von
Jhering remarked in response to Savigny’s asssertion that legal form requirements prevented
rash contract formations and advanced the goal of ‘well-considered deliberation’, the extent
to which this goal can be achieved depends upon the nature of the form required: ‘This [scil.:
Savigny’s argument] is certainly true for those forms that lead to a certain delay . . .’.127
Other form requirements, however, can be complied with in such a brief period that ‘the
concomitant delay [is] much too short to leave a party, who had promised something in
excitement or a hurry, time to calm down or to deliberate.’128 Similarly, Hillman reports that
‘some distrust NOM clauses because they suspect that a writing does not necessarily
promote the cautionary . . . function traditionally associated with legal formalities. In our
hurried business world, people often sign writings without reading or contemplating
them.’129

125
World Online v I-Way [2002] EWCA Civ 413 [17].
126
Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280, 1286: ‘[I]t was natural [for the drafters
of the UCC] to give the parties some means of providing a substitute for the cautionary and evidentiary
function that the requirement of consideration provides; and the means chosen was to allow them to exclude
oral modifications’; Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 450 (1988)
(mentioning the protection against ‘inadvertent or unwise oral adjustment’ even before the prevention of
‘fraudulent or mistaken claims’).
127
von Jhering, Geist des römischen Rechts II.2, 497.
128
Ibid.
129
Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 451 (1988). In the omitted part of the
sentence, Hillman also claims that NOM-clauses need not promote the evidentiary function either, later
explaining that ‘wrongdoers can easily produce “expert” forgeries.’ (ibid). I fail to see the point of this
addition; while no kind of evidence is immune against tampering, there can be no doubt that it is easier and
less risky to manufacture oral evidence than written evidence. Similarly, I do not understand the argument that
‘a NOM clause may even promote fraud by encouraging a party wrongfully to deny the existence of an oral
agreement to adjust a contract term’ (ibid, 452): If the NOM-clause is enforced, there is no need even for the
most notorious liar to deny any oral agreement (except to avoid the exceptions to the enforcement of NOM-
clauses, such as waiver or estoppel; but if anything, this is an argument against the exceptions, not against the
enforcement of the rule).
34
 Nevertheless, von Jhering pointed out that the delay required to comply with the
form requirement is not its only cautionary aspect: having to sign a writing will at least alert
the parties to the fact that they are leaving the purely social or conversational context and
that they are entering legal domain; this will put them on their guard.130 This latter aspect of
the cautionary function, which overlaps with the channelling function, has a role to play
even in the commercial context. A conversation on the golf course may be intended as a
binding promise or as banter. Banter may turn into business negotiations. If one of the
parties then suggests the formalization of the outcome of such a conversation, the other
party will be in no doubt as to the first party’s intentions.

A Check on Agents?
It has been suggested that yet another function of NOM-clauses is to protect the principal
against ‘thoughtless or overzealous agents’.131 This highlights the relationship between the
law of agency and NOM-clauses. Where the agent has no authority to modify the written
agreement – especially where the NOM-clause is interpreted as denying the agent this
authority –,132 the written contract may prevail over oral concessions made by the agent
even in jurisdictions where NOM-clauses as such are given no effect. While ‘the parties’
can ‘make and unmake’ their contracts, not every agent can. This can even be used to
simulate artificially the effects of a desired NOM-clause in these jurisdictions by letting an
agent without the authority to modify the written contract deal with the day-to-day
negotiations with the other side.133 However, this is not a benefit of the NOM-clause as such
but a result of the law of agency. Accordingly, the effectiveness of such a strategy may be
limited by implied or apparent authority.

Summary of Benefits
The parties therefore have a legitimate interest in establishing a NOM-clause, be it for
evidentiary, channeling or cautionary purposes. In economic terms, if the legal system

130
von Jhering, Geist des römischen Rechts II.2, 497–8; LL Fuller, ‘Consideration and Form’, (1941) 41
Columbia Law Review 799, 800: ‘inducing the circumspective frame of mind’. Fuller knew, cited and quoted
von Jhering’s work.
131
JK Brooks, ‘Parol Modification and the Statute of Frauds: Fitting the pieces together under the Uniform
Commercial Code’ (1999) 21 Campbell Law Review 307, 314; for more detail and further references see RB
Ahdieh, ‘The Strategy of Boilerplate’ 104 Michigan Law Review 1033, 1041 in fn 30 (2006)
132
This may be the case in particular in the case of so-called ‘Home-Office Approval’ clauses, which require
any modifications to be made or approved by the central back office.
133
Home-Office Approval clauses are an attempt to do exactly this. While this may simulate the effect of an
effective NOM-clause, it is a costly device: An agent with limited authority has to be employed.
35
enforces NOM-clauses, the parties will increase their transaction costs for future contract
modifications, but this allows them to reduce:

(i) the costs for monitoring their agents;

(ii) the probability of conflicts arising and thus the expected costs of such
conflicts;
(iii) the danger of opportunism by the other party;
(iv) the costs of taking oral evidence in potential future litigation or
arbitration; and
(v) the uncertainty about the outcome of such litigation or arbitration;
this increases the utility for risk-averse parties.

Policy Arguments Against the Enforcement of NOM-clauses

In contrast, it is difficult to find convincing policy reasons for denying NOM-clauses any
effect.134 Certainly, enforcing the NOM-clause has its costs as well as its benefits. Jhering
summarized the disadvantages of (statutory) form requirements as their ‘dangerousness’ and
their ‘inconvenience’: Form requirements are dangerous because of the severe consequences
in case of non-compliance; and compliance with them may be inconvenient in the day-to-
day business.135 The dangerousness is particularly problematic where one party, knowing of
the form requirement, opportunistically lures an unwitting counter-party into relying on an
invalid modification: ‘Under a rule of formality, the honest, but commercially inexperienced
man is more at a disadvantage vis-à-vis the clever and ruthless business man than under a
rule of informality, because he who knows how to use the form will lay a trap for the
unwary . . .’136
While these objections can be advanced not only against statutory form requirements,
but also against NOM-clauses, they have less force in the context of NOM-clauses.

- First, with regard to the ‘inconvenience’ of the form requirement: it was the parties
who agreed to the NOM-clause, thus revealing their preference for incurring the

134
Martinsville Nylon Employees Council Corp v National Labor Relations Board, 969 F.2d 1263, 1268 (DC
Cir. 1992) (‘virtually nothing to commend . . . the common law rule’).
135
von Jhering, Geist des römischen Rechts II.2, 480–9.
136
von Jhering, Geist des römischen Rechts II.2, 481.
36
 cost of compliance with the form, and so implicitly revealing their calculation that
this cost is offset by the (evidentiary and/or cautionary) benefits of the form.137
- Secondly, the ‘dangerousness’ of the form requirement crucially depends on the lack
of knowledge about the form requirement. The danger that the form requirement is
overlooked is greater in the case where a statute, possibly unknown to the parties,
establishes a form requirement than in the case where the parties themselves have
established the form requirement. Admittedly, the danger may be nearly as great
where the form requirement is hidden away in the standard terms and conditions; but
for these cases, one can apply special rules.138
- Thirdly, in so far as the concern is the opportunistic behaviour of businessmen
against consumers, this may be addressed by consumer protection rules. One could
consider restricting the use of NOM-clauses in standard terms and conditions, and/or
enforcing NOM-clauses only in contracts ‘between merchants’ or ‘between
businesses’.139
- Fourthly, in so far as even businesses may fall prey to such opportunistic conduct
and in so far as one allows the use of NOM-clauses by or against consumers, the
alternative to not enforcing the NOM-clause at all is to establish certain exceptions
only for the cases in which such opportunistic conduct can be shown to exist.140

The enforcement of NOM-clauses has its costs and its benefits, and there is no conclusive
argument for either side. One commentator, who recognized and frankly admitted to this
inconclusive nature, opined that ‘[t]he answer to this question depends on comparing the
frequency and costs of parties mistakenly thinking a NOM clause benefits them with the
frequency and benefits of parties correctly including a NOM clause.’141 While this has
utilitarian appeal, I am not certain that it is necessarily true in a libertarian system: Even if
many parties mistakenly include a NOM-clause, this does not mean that all parties should to
be deprived of the opportunity of using enforceable NOM-clauses.142
In addition, one has to be careful not to overestimate the ‘costs of parties mistakenly
thinking a NOM clause benefits them’. As von Jhering pointed out:

137
cf the quotes accompanying nn 58, 124.
138
Below 9.
139
Below 9.
140
Below 8.
141
Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 452 (1988).
142
See the quote above n 124 (in particular: ‘The drawback of the common law rule is that it denies to all
contracting parties, no matter how sophisticated, the ability to decide for themselves . . .’).
37
 With formalism, it is as with many other institutions – the oppressive effects are felt
by everybody, the beneficial effects by nobody, because the latter are only of a
negative kind, i.e. averting trouble. The one case in which the disadvantages of a
form requirement become apparent . . . will be talked about more widely than the
thousands of cases in which the course of events was a regular one and the form has
fulfilled its beneficial function.143
And later:
One supposes unconsciously that he [scil. the person invoking the invalidity due to
lack of form] really intended to legally bind himself, and that he only negligently
forgot to comply with the required form. And yet, the opposite is also possible,
namely that the omission of the form had its reason in the intention to avoid
becoming legally bound . . .144
And, to invoke von Jhering one last time:
If formalism entails the danger that somebody who really had the intention to bind
himself is acquitted from liability because of non-compliance with the form
requirement, then the system of informality entails the opposite danger, namely that
somebody who did not have this intention is held liable against his true will; which
of these dangers is the more problematic one will not be open to doubt.145

Going Back to the Good Old Times?

It has also been suggested that the development and enforcement of NOM-clauses is a
relatively new invention, and that it may be preferable to go back to the good old times of
the common law.146 I do not subscribe to Henry Ford’s view that ‘history is more or less
bunk’. But if one does want to go down this historically-minded route, then one also has to
realize that for several centuries, and in the United States well into the twentieth century, the
parties had the option of producing the effect of an enforceable NOM-clause by drafting an

143
von Jhering, Geist des römischen Rechts II.2, 480.
144
Ibid, 492.
145
Ibid, 498. If the reader should still be doubtful: In the first case the party can protect itself by being vigilant
and complying with the form requirement; in the second case, the party has no such possibility. And, to
reiterate, in contrast to the statutory form requirements discussed by von Jhering, in the case of contractual
NOM-clauses the parties have even imposed the form requirements onto themselves.
146
ie, ‘the common law ban on NOM clauses’. Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal
449, 463 (1988), pointing out that the common law position ‘hardly wreaked havoc on commercial
transactions. . . [W]e should preserve the status quo ante’.
38
instrument under seal; such an instrument could then only be modified or rescinded by an
instrument under seal.147

5. Limitations to the Effectiveness of NOM-clauses: an Overview

Even where NOM-clauses are generally enforced, subsequent oral agreements are not
always completely without effect. Four cases can be distinguished: The first case is that the
subsequent oral agreement is considered to be a rescission of the original agreement and the
formation of a new agreement on new terms – in short, a novation (below 6). The second
case is that the informal agreement is considered to be an independent, collateral agreement,
existing alongside the (unmodified) original agreement (below 7). The third case is that one
of the parties has intentionally relinquished a known right, and/or that the other party has
relied on an informal agreement so that it would be inequitable for the other party to go back
on this agreement (below 8). A fourth consideration is the protection against ‘surprising’
NOM-clauses in the small print of standard terms and conditions. Where NOM-clauses are
not individually negotiated, some jurisdictions provide for special rules with regard to their
effectiveness (below 9).

6. Informal Rescission, Novation, and NOM-clauses

Where the parties to a contract containing a NOM-clause purport to modify the contract
orally, this purported modification may sometimes be interpreted as being a rescission148 of

147
See above n 21. For the equivalent rule under Roman law see von Jhering, Geist des römischen Rechts II.2,
626 et seq.
148
In keeping with the US American terminology (Restatement (Second) § 283; see also UCC § 2-209(2)), I
use the term rescission here to denote the mutually agreed ending of the contract ex nunc (unless the parties
agreed to ab initio effects). In the CISG- and UNIDROIT-context, I use ‘termination’ to denote the same thing
in accordance with the respective provisions on NOM-clauses. UCC § 2-106(3) reserves ‘termination’ for
unilaterally ending a contract ‘pursuant to a power created by agreement or law . . . otherwise than for its
breach’; executory obligations are ‘discharged’ but ‘any right based on prior breach or performance survives’.
Similarly, UCC § 2-106(4) reserves ‘cancellation’ for the ending of the contract for the other party’s breach.
Outside the UCC-context, the terminology, both between and within jurisdictions, differs widely: Some
authors reserve ‘termination’ for ending a contract following a breach, others reserve ‘rescission’ to voidable
contracts or to the vitiation of the contract ab initio (eg Wilken, Waiver, Variation and Estoppel (2nd edn,
Oxford: Oxford University Press, 2002) para 2.33, 2.35); some use the term ‘discharge’ (ibid, para 2.34; this
term is reserved by others for the automatic discharge, such as the one following a frustrating event, or for the
complete cessation of ‘any legal effect’ of a contract as opposed to termination, which leaves contractually
agreed post-contractual duties in place, Samuel (Professionally known as Seal) v Wadlow [2007] EWCA Civ
155, 2007 WL 555756 [43]), ‘cancellation’, ‘abrogation’ or ‘abandonment’. While Art 29(2) CISG and Art
2.1.18 of the UNIDROIT Principles use the term ‘termination’, the Restatement (Second) of Contracts § 283
and UCC § 2-209(2) use the term ‘rescission’ (which, according to comment 3 to UCC § 2-209, includes
‘abandonment’; similarly comment a. to Restatement (Second) § 283). § 15-301 New York General
Obligations Law (text below n 170) uses both ‘discharge’ and ‘termination’. In the drafting history of the
CISG, the predecessor of Art 29, Art 27 of the 1978-Draft, used the term ‘abrogation’ before it was changed to
‘termination by agreement’ (cf the CISG Secretariat Commentary on the 1978 Draft, Article 27 para 1,
39
the old contract coupled with the conclusion of a new contract – the purported modification
could result in a ‘novation’.
Unless a statute provides otherwise, a NOM-clause can arguably not prevent the
informal rescission of the contract as a whole. This can be derived from an analogy to the
situation where the parties try to modify a contract whose formation is subject to a statutory
form requirement.

Modification v Novation of Contracts Subject to a Statutory Form Requirement

Where the parties attempt to modify a contract whose formation is subject to a statutory
form requirement, this may lead to one of three results:

- Either, the parties really want only to modify the original contract, in which case the
question is whether the modification is valid or invalid (or unenforceable). This
depends on whether the form requirement extends to the modification in question.
Usually, modifications of material terms of contracts that are subject to a statutory
form requirement necessitate compliance with the form requirement. If they are not
in the form required, the law can react in one of two ways: either the entire contract
becomes invalid for lack of compliance with the required form149 or the contract
between the parties remains unmodified, but valid with its original content.150
- Alternatively, the parties may intend a major departure from the original contract: If
this is the case, then one could interpret the attempted modification as the rescission

pointing to the discussion between Mr Medvedev, USSR, and Professor Farnsworth, USA, in Official Records,
272).
149
In Germany, if there is a statutory form requirement and the parties try to modify the contract, the general
rule is that certainly the modification itself, but usually also the entire contract, becomes void for lack of form
if the statutory requirement requires the entire contract to be in a certain form, unless the parties’ intention
would have been to sever the invalid modification (§139 of the German Civil Code). In some instances, the
courts have made an exception for minor modifications, see eg Bundesgerichtshof 2 July 1975, case VIII ZR
223/73, 65 Entscheidungen des Bundesgerichtshofs in Zivilsachen 49, [1975] Neue Juristische Wochenschrift
1653; Bundesgerichtshof 27 March 1968, case VIII ZR 71/66, 50 Entscheidungen des Bundesgerichtshofs in
Zivilsachen 39, [1968] Neue Juristische Wochenschrift 1229, 1230; against such exceptions: Hertel in
Staudingers Kommentar § 125 paras 76–7. The details are of no importance for present purposes.
150
This is usually the case in England, eg McCausland v Duncan Lawrie Ltd [1997] 1 WLR 38 (CA); see also
Noble v Ward (1866–67) LR 2 Ex 135 (Exch. Ch.), as interpreted by Lord Dunedin in Morris v Baron [1918]
AC 1, 26–7. In United Dominions Corp (Jamaica) v Shoucair [1969] 1 AC 340 (PC), Lord Devlin
distinguished (at 347) according to whether the statute made the variation ‘void’ or only ‘unenforceable’: If the
statute made the variation ‘void and of no effect, there would be no problem at all. An attempt at changing the
original contract would have failed altogether and so left it quite untouched.’ If the statute made the variation
only unenforceable (as it did in the case before the Privy Council), then a distinction was said (at 348) to be
necessary ‘based on the intention of the parties’: ‘If the new agreement reveals an intention to rescind the old,
the old goes, and if it does not, the old remains in force and unamended.’
40
 of the original contract – which is usually not subject to a form requirement –,151
coupled with the attempt to form a new contract on the new terms. As the formation
of the new contract is subject to the statutory form requirement, the overall result is
that the parties may have rescinded the old contract without forming a valid new
contract.152

Clearly, the various interpretations of the attempted modification have very different
consequences. Despite the gravity of these consequences, there is no bright-line test for the
distinction between modifications and novations. Instead, the distinction is said to be ‘one of
degree’153 and ‘may not be an easy one to answer’.154 While the decision ultimately turns on
the (manifested) intent of the parties,155 the change ‘is a rescission if it alters the original
contract in some essential way’.156

Modification and Novation in the Case of NOM-clauses

If even statutory form requirements cannot prevent the contract from being rescinded
informally, a NOM-clause in itself (as a ‘private Statute of Frauds’) is unlikely to provide
greater protection against an informal rescission of the contract containing the NOM-
clause.157 If the attempted modification is interpreted as being a novation, the original

151
In England: Morris v Baron [1918] AC 1 (HL); in Germany: Hertel, Staudingers Kommentar § 125 para
80–3; M Löwisch, Staudingers Kommentar (2005) § 311 para 78 (but see the qualifications in paras 79, 80);
Ellenberger in Palandt § 125 para 10; Grüneberg in Palandt § 311 para 7; J Kindl in Erman § 311 para 14.
152
Morris v Baron [1918] AC 1 (HL).
153
E Peel, Treitel The Law of Contract (12th edn, London: Sweet & Maxwell, 2007) § 5-030.
154
McKendrick in Chitty on Contracts § 22-034.
155
England: Morris v Baron [1918] AC 1, 19 (HL); United Dominions Corp (Jamaica) v Shoucair [1969] 1
AC 340, 348 (PC) (see the quote above n 150); McKendrick in Chitty on Contracts § 22-034; Wilken, Waiver,
Variation and Estoppel para 2.31–2. California: Fanucchi & Limi Farms v United Agri Products, 414 F.3d
1075, 1082 (9th Cir 2005); Alexander v Angel, 37 Cal.2d 856, 860, 236 P.2d 561 (1951). Germany:
Bundesgerichtshof 1 October 2002, case IX ZR 443/00 [2003] Neue Juristische Wochenschrift 59;
Bundesgerichtshof 14 November 1985, case III ZR 80/84 [1986] Neue Juristische Wochenschrift 1490; see
also Bundesgerichtshof 30 September 1999, case IX ZR 287/98 [1999] Neue Juristische Wochenschrift 3708,
3709; M. Löwisch in Staudingers Kommentar (2005) § 311 para 71; Kindl in Erman § 311 para 11.
156
Peel Treitel Law of Contract, § 5-030. For English law: British and Beningtons v North Western Cachar
Tea [1923] AC 48, 62, 68 (no rescission if it ‘does not go to the very root of the contract’ (Lord Sumner)). For
Californian law, see Alexander v Angel, 37 Cal.2d 856, 861, 236 P.2d 561 (1951) (asking if the changes were
‘sufficiently substantial’ or ‘drastic’ to show an extinguishment of the old obligation and replacement of the
new one); see also Fanucchi & Limi Farms v United Agri Products, 414 F.3d 1075, 1082-8 (9th Cir 2005) (but
see Judge Beezer’s concurrence, ibid, 1090–4). For German law: cf Löwisch in Staudingers Kommentar § 311
para 73. In German law, in cases of doubt, there is an evidential presumption in favour of a mere modification,
see, eg Bundesgerichtshof 1 October 2002, case IX ZR 443/00 [2003] Neue Juristische Wochenschrift 59;
Bundesgerichtshof 14 November 1985, case III ZR 80/84 [1986] Neue Juristische Wochenschrift 1490;
Löwisch, ibid, para 74; Kindl in Erman § 311 para 11.
157
cf Restatement (Second) of Contracts § 283, comment b: ‘Even a provision of the earlier contract to the
effect that it can be rescinded only in writing does not impair the effectiveness of an oral agreement of
rescission. In the absence of a statute, such a self-imposed limitation does not limit the power of the parties
41
contract will be rescinded in its entirety – including the NOM-clause. In contrast to the
situation where the form requirement is statutory, the formation of the new contract is not
subject to any form requirement, provided the form requirement resulted exclusively from
the NOM-clause in the original – and now rescinded – contract. Once the attempted
modification is interpreted as being a novation (rather than merely an attempt to modify the
original contract that is invalid because of the NOM-clause), the original contract is
rescinded and replaced by a new contract on the new terms. The overall effect of a novation
is here identical to that of a modification.
The consequence is that even in some jurisdictions that generally enforce NOM-
clauses, the NOM-clause does not protect against a novation. Because the distinguishing
criterion between a ‘mere modification’ and a ‘novation’ is whether the alteration of the
original contract is ‘essential’, 158 the somewhat perverse consequence is that in such
jurisdictions a NOM-clause will protect the parties against minor oral modifications, but not
against essential changes.
The problem is illustrated by the Californian Fanucchi case, decided in 2005 by the
US Court of Appeals for the Ninth Circuit.159 Fanucchi & Limi Farms (Fanucchi) owed
more than one million dollars to United Agri Products (United) under a loan agreement
containing a NOM-clause. Fanucchi claimed that when it had encountered difficulty in
repaying the loan and considered filing for bankrupcy, a representative of United had orally
agreed to the following terms: United would subordinate its loan to new loans from other
lenders; any proceeds not used to service new loans would be split 60:40 between United
and third-party creditors of Fanucchi; and if after five years Fanucchi had paid the United
loan down to ‘$300,000 or $400,000’, United would forgive the outstanding debt.
The District Court had granted summary judgment to United, arguing that NOM-
clauses are enforceable under § 1698(c) of the Californian Civil Code,160 and the alleged
agreement between Fanucchi and United’s representative had not been reduced to writing.
The Court of Appeals agreed that the NOM-clause prevented an oral modification. It noted,
however, that § 1698(d) of the Californian Civil Code explicitly provides that ‘[n]othing in

subsequently to contract.’ As this passage indicates, a statutory provision giving effect to NOM-clauses can at
the same time require that a rescission conform to the form established in the NOM-clause (see below,
describing NY Gen Oblig L § 15-301, UCC § 2-209, CISG Art 29(2) and UNIDROIT Principles 2.1.18). If a
statute enforces NOM-clauses against rescission, then the NOM-clause may indeed have greater effect than
most statutory form requirements.
158
See above n 156. Note that some reserve the term ‘novation’ to the change of parties to a contract. I use it in
the more general meaning, the replacement of the original contract with a new one.
159
Fanucchi & Limi Farms v United Agri Products, 414 F.3d 1075 (9th Cir 2005).
160
For the wording of § 1698(c) Cal Civil Code see n 64.
42
this section precludes in an appropriate case the application of rules of law concerning . . .
oral novation and substitution of a new agreement . . .’ The Court of Appeals therefore went
on to consider whether the changes alleged by Fanucchi were sufficiently ‘drastic’ to be
considered an oral novation, and concluded that they were. In consequence, the Court of
Appeal reversed the grant of summary judgment and remanded the case for trial – a result
the parties arguably had wanted to avoid by including the NOM-clause.161 If Fanucchi had
only claimed that United had agreed to late payment, this would presumably not have been
sufficiently drastic to constitute a novation and would therefore have amounted only to an
invalid modification – and Fanucchi would have lost at summary judgment stage. It pays to
be bold.
The Fanucchi-example from the West coast of the United States demonstrates that a
statutory provision giving effect to NOM-clauses should at the same time protect the parties
against oral novations, lest the NOM-clause be ineffective in cases where the alleged
‘modification’ is particularly drastic.
On the East coast of the United States, namely in New York, this recognition had led
to legislative changes long before.162 As described above, the New York Court of Appeal
had originally held NOM-clauses to be ineffective under New York common law in the
1919-decision Beatty v Guggenheim Explorations.163 At the time Beatty was decided, it was
still possible to create an effect similar to that of a NOM-clause by contracting under seal,
because contracts under seal could not be modified by parol.164 Later, in 1941, when the seal
was abolished in New York, statutory provisions sought to ensure the effectiveness of
NOM-clauses.165 While these provisions already protected against an oral discharge – and
therefore arguably would have avoided the problems encountered by the Fanucchi-Court in
California –, they did not anticipate the constellation in Green v Doniger, a 1949 case in
which the contract contained a NOM-clause, but also permitted the parties to terminate the

161
A similar problem arose in Germany, when a party (as in Fanucchi) claimed that the other party had
forgiven its debt: Kammergericht 18 August 2005, case 8 U 106/04 [2005] BeckRS 10787, [2006] OLG
Report KG 86 (part II.2 of the judgment). As in Fanucchi, the German court decided that the forgiveness of
debt (under § 397 of the German Civil Code) was not subject to the NOM-clause. In the end, this did not
matter, because the person acting for the party allegedly forgiving the debt had lacked authority.
162
For a concise account of the development described in the text, see Israel v Chabra, 12 N.Y.3d 158, 163–7,
906 N.E.2d 374, 377–80 (NY 2009).
163
Above n 33.
164
Cammack v JB Slattery & Brothers, 241 N.Y. 39, 45–6, 148 N.E. 781 (1925). For a similar position for
deeds under the English common law, see n 21.
165
Real Property Law § 282 and Personal Property Law § 33-c, L. 1941, ch 329, §§ 4, 5 (the predecessor
provisions to Gen Oblig L § 15-301(1), the text of which is cited in n 170).
43
contract unilaterally on written notice.166 The plaintiff in Green claimed that while the
original written contract had not made provision for a bonus payment, he and his employer
had subsequently orally agreed on a bonus. The New York Court of Appeals held that the
NOM-clause was effective under the statute, at least with regard to modifications. However,
the Court went on to consider that the termination-on-notice provision in the contract
allowed the unilateral termination of the contract, and concluded that the parties had thereby
made special provision for the ‘discharge’ and that accordingly the parties intended the
statutory provision to be inapplicable with regard to discharges.167 As a consequence, it was
possible that the employee had unilaterally and orally terminated the contract and that the
employer had waived the requirement of the written form requirement for the notice; and
that after having thus rescinded the original contract, they had substituted an oral contract on
the new terms including the bonus payment. Accordingly, the lower court’s decision to
grant a motion to dismiss was reversed. Once again, the ‘freedom of contract’ argument had
defeated the NOM-clause.168 And once again, the New York legislature intervened to close
the gap: In 1952, it amended the predecessor statute of § 15-301 of the General Obligations
Law. 169 Today, § 15-301 of the New York General Obligations Law seeks to give
comprehensive effect to NOM-clauses by foreclosing each of the many ways to circumvent
a NOM-clause.170

166
Green v Doniger, 300 N.Y. 238, 90 N.E.2d 56 (1949).
167
300 N.Y. 238, 244–5, 90 N.E.2d 56, 59: ‘The statutory clause, which when inserted in a contract subjects it
to the control of subdivision 1 of section 33-c, is a simple clause prohibiting oral change. . . . Not only is oral
change prohibited, but there may be no change, modification, or discharge, in whole or in part, without a
writing signed by the party to be charged. Whether the statute shall apply to a given contract depends upon the
intent of the parties as expressed by the inclusion or exclusion, in some form, of the statutory clause
prohibiting oral change. In the present case such a clause in the written contracts evidences the intent of the
parties that the statute shall have at least a limited application. The scope of that application, however, appears
to be intentionally limited to modifications or changes to the exclusion of discharges, since the parties have set
up their own less formal procedure for discharge or termination.’ (emphasis added).
168
Judge Fuld dissented: ‘To hold that one party may accomplish a change in the amount of compensation to
be paid to him under a written contract containing such a clause by the simple expedient and device of calling
the change an abandonment, would be to annul the statute and render its operation nugatory.’ 300 N.Y. 238,
247, 90 N.E.2d 56, 61.
169
See L 1952, ch 831, and Israel v Chabra, 906 N.E.2d 374, 379 (NY 2009).
170
§ 15-301 of the NY General Obligations Law provides in full:
1. A written agreement or other written instrument which contains a provision to the effect that it
cannot be changed orally, cannot be changed by an executory agreement unless such executory
agreement is in writing and signed by the party against whom enforcement of the change is sought or
by his agent.
2. A written agreement or other written instrument which contains a provision to the effect that it
cannot be terminated orally, cannot be discharged by an executory agreement unless such executory
agreement is in writing and signed by the party against whom enforcement of the discharge is sought,
or by his agent, and cannot be terminated by mutual consent unless such termination is effected by an
executed accord and satisfaction other than the substitution of one executory contract for another, or
is evidenced by a writing signed by the party against whom it is sought to enforce the termination, or
by his agent.
44
 If ‘informal novations’ are given effect even in the presence of enforceable NOM-
clauses, then hair-splitting distinctions between rescission and modification, based on the
‘essence of a contract’ will be required. 171 The NOM-clause will prevent minor
modifications, but not – alleged – essential changes to the contract. What is more, a party
aspiring to go to trial in order to introduce oral evidence can circumvent the NOM-clause by
claiming a particularly drastic ‘modification’ of the original contract, forcing the court to
hear oral evidence. Even though the court will require ‘clear and convincing evidence’ to
find an oral novation, the parties’ intention to prevent the case turning on – by its nature
unreliable – witness testimony will have been thwarted. If a legislator wants to avoid this
result, then it should make certain that NOM-clauses protect against informal novations as
well.

7. Collateral Agreements and NOM-clauses

A problem similar to that of NOM-clauses being undermined by a novation can arise where
a party does not claim that the informal agreement modified the written contract containing
the NOM-clause, but that the informal agreement is a collateral contract, standing
independently beside the original contract.172

3. a. A discharge or partial discharge of obligations under a written agreement or other written

instrument is a change of the agreement or instrument for the purpose of subdivision one of this
section and is not a discharge or termination for the purpose of subdivision two, unless all executory
obligations under the agreement or instrument are discharged or terminated.
b. A discharge or termination of all executory obligations under a written agreement or other written
instrument is a discharge or termination for the purpose of subdivision two even though accrued
obligations remaining unperformed at the date of the discharge or termination are not affected by it.
c. If a written agreement or other written instrument containing a provision that it cannot be
terminated orally also provides for termination or discharge on notice by one or either party, both
subdivision two and subdivision four of this section apply whether or not the agreement or other
instrument states specifically that the notice must be in writing.
4. If a written agreement or other written instrument contains a provision for termination or discharge
on written notice by one or either party, the requirement that such notice be in writing cannot be
waived except by a writing signed by the party against whom enforcement of the waiver is sought or
by his agent.
5. If executed by an agent, any agreement, evidence of termination, notice of termination or waiver,
required by this section to be in writing, which affects or relates to real property or an interest therein
as defined in section 5-101 in any manner stated in subdivisions one or two of section 5-703 of this
chapter shall be void unless such agent was thereunto authorized in writing.
6. As used in this section the term “agreement” includes promise and undertaking.
171
For a similar distaste for the distinction between modifications and rescission, see Samuel (Professionally
known as Seal) v Wadlow [2007] EWCA Civ 155, 2007 WL 555756 [34]–[46]. See also Wilken, Waiver,
Variation and Estoppel paras 2.30–2.38 (proposing that all attempts at variation should be treated as an
implied discharge).
172
England: AS Klaveness Chartering v Pioneer Freight Futures [2009] EWHC 3386 (Comm), 2009 WL
4872647 [23]: ‘PFF would contend that, if there was such an agreement it amounts to an amendment,
modification or waiver of the Master Agreement and must, in order to be effective, be in writing . . . There is
authority that a contract which is collateral to a contract which is required to be in writing does not itself have
to be in writing. Angell v Duke (1875) LR 10 QB 174; City of Westminster Properties v Mudd [1959] Ch 129;
45
 As in the case of novations, the distinction between modifications and collateral
contracts will require difficult value judgments. The distinguishing criterion is usually
whether the subject matter of the informal agreement is independent in that it does not
qualify or contradict the original contract.173 On the one hand, this criterion is at least

Record v Bell [1991] 1 WLR 853; Chitty on Contracts (30th ed, 2008), paragraphs 12-103 and 22-036’. Also
cf Rix LJ, quoted in World Online v I-Way [2002] EWCA Civ 413 [7]: ‘That agreement [scil. the modification
in question] can either be treated as an oral agreement varying the original agreement, or as a free-standing
contract, and I know nothing in case law that prevents such an event having effect.’
For the US, see eg California Civil Code § 1698(d), which in full provides: ‘(d) Nothing in this
section precludes in an appropriate case the application of rules of law concerning estoppel, oral novation and
substitution of a new agreement, rescission of a written contract by an oral agreement, waiver of a provision of
a written contract, or oral independent collateral contracts.’ (emphasis added).
173
For California, see Davidson v Conocophillips, 2009 WL 2136535 (ND Cal. 2009), slip copy at 5, citing
Dobbins v Horsfall, 58 Cal App.2d 23, 29, 136 P.2d 35 (1943) and Malmstrom v Kaiser Aluminum &
Chemical Corp, 187 Cal App.3d 299, 318, 231 Cal Rptr 820 (1986).
The English position is not entirely clear. Older cases dealing with antecedent or contemporaneous
oral agreements accompanying written contracts tended to use the same criterion, ie whether the agreement
was independent in that it did not contradict or qualify the main contract. See Henderson v Arthur [1907] 1 KB
10, 12–3, where Collins MR stated: ‘[T]o admit evidence of such an agreement . . . would be to violate one of
the first principles of the law of evidence; because . . . it would be to substitute the terms of an antecedent
parol agreement for the terms of a subsequent formal contract under seal dealing with the same subject matter.
I do not see how . . . the covenant in the lease and the antecedent parol agreement can co-exist . . . [I]t was not
a merely collateral agreement, but provides in another and contradictory manner for doing what was
subsequently provided for by the lease.’; Goldfoot v Welch [1914] 1 Ch 213, 218: ‘Having arrived at what I
think to be the true construction of the [scil.: written] agreement, I am of opinion that no evidence to contradict
it was admissible’ (Eve J); see also Angell v Duke (1874–75) LR 10 QB 174 (where an oral agreement to
repair and supply furniture was found to be ‘something antecedent and collateral . . . that is, a separate
agreement entered into in order to induce the intended tenant to accept the tenancy’).
Some later cases, however, allow qualifications and contradictions in the case of such antecedent or
contemporaneous oral agreements. Heilbut, Symons & Co v Buckleton [1913] AC 30 took an intermediate
position; see Lord Moulton’s speech at 48: ‘It is evident, both on principle and on authority, that there may be
a contract the consideration for which is the making of some other contract. . . It is collateral to the main
contract, but each has an independent existence, and they do not differ in respect of their possessing to the full
the character and status of a contract. But such collateral contracts must from their very nature be rare. . . .
[T]he more natural and usual way of carrying this out would be by . . . modifying the main contract and not by
executing a concurrent and collateral contract. Such collateral contracts, the sole effect of which is to vary or
add to the terms of the principal contract, are therefore viewed with suspicion by the law. They must be proved
strictly. Not only the terms of such contracts but the existence of an animus contrahendi on the part of all the
parties to them must be clearly shewn. Any laxity on these points would enable parties to escape from the full
performance of the obligations of contracts unquestionably entered into by them and more especially would
have the effect of lessening the authority of written contracts by making it possible to vary them by suggesting
the existence of verbal collateral agreements relating to the same subject-matter.’ (emphasis added). Later
cases went further, see Couchman v Hill [1947] KB 554, 558; J Evans & Son v Andrea Merzario Ltd [1976] 1
WLR 1078, 1081 (where Lord Denning MR said, referring to Heilbut: ‘[W]e have a different approach to
collateral contracts.’); also cf Brikom Investments v Carr [1979] QB 467 (where the Court agreed on the
outcome, but where there was disagreement whether this resulted from equitable estoppel, waiver and/or a
collateral agreement); Record v Bell [1991] 1 WLR 853 (Ch); see also AG Guest in Chitty on Contracts § 12-
103.
At least one distinguished commentator has suggested that the same rules apply to the category with
which we are concerned here, namely subsequent collateral contracts. If that were the case, subsequent oral
collateral contracts could contradict the written contract. See McKendrick in Chitty on Contracts § 22-036.
This view finds some support in dicta from AS Klaveness and World Online v I-Way (quoted above in n. 172).
With respect, I would prefer to distinguish the cases of antecedent or contemporaneous oral agreements
described above from subsequent oral collateral contracts. For subsequent collateral contracts, one should
retain the distinguishing criterion of whether the oral agreement contradicts the original contract. What was
really at issue in the cases of contemporaneous contradictory collateral agreements cited above was whether
46
the ‘written’ agreement was the whole (ie an integrated) agreement, or whether it was a partially written,
partially oral agreement (in which case extrinsic evidence is admissible; see generally Peel in Treitel Law of
Contract § 6-013). That this was the question becomes particularly clear in Lord Justice Roskill’s judgment in
J Evans & Son v Andrea Merzario (at 1083): ‘That phrase [scil.: collateral oral warranty] is normally only
applicable where the original promise was external to the main contract, that main contract being a contract in
writing, so that usually parole evidence cannot be given to contradict the terms of the written contract . . . But
that doctrine . . . has little or no application where one is not concerned with a contract in writing . . . but with
a contract which . . . was partly oral, partly in writing, and partly by conduct. In such a case the court does not
require to have recourse to lawyer’s devices such as collateral oral warranty . . . The court is entitled to look at
and should look at all the evidence from start to finish in order to see what the bargain was that was struck
between the parties.’; and with this, Geoffrey Lane LJ agreed (at 1084): ‘This was not a collateral contract in
the sense of an oral agreement varying the terms of a written contract.’ Similarly, Lord Justice Roskill in
Brikom Investments stated: ‘This case seems to me to fall within the principle laid down by this court in De
Lassalle v Guildford [1901] 2 K.B. 215 . . .: “. . . It appears . . . clear that the lease did not cover the whole
ground, and that it did not contain the whole of the contract between the parties” ’, and Cumming Bruce, LJ
agreed with Roskill, LJ. To this argument one could object that in Brikom Investments the contract contained
an entire agreements clause (quoted by the Court at 480); but I take the Court to hold that, given that such
clauses establish merely a rebuttable presumption that the written agreement contains the whole agreement, the
presumption had been rebutted. At any rate, the decision in Brikom Investments can be explained as an
application of waiver (Roskill and Cumming Bruce, LJJ) or estoppel (Lord Denning, MR); the issue of a
collateral contract was not extensively discussed, but only added as an afterthought by Lord Denning, MR (at
485).
An additional factor that may have influenced the courts in some of the cases giving effect to the
contemporaneous oral agreement was that the written ‘contract’ consisted in standard terms and conditions (eg
Couchman v Hill; J Evans & Son). The reason for allowing evidence of the contemporaneous oral agreement
may have been the desire to let individually negotiated terms trump the standard terms (similarly to § 305b of
the German Civil Code and equivalent rules, see below n 207). This desire surfaces in Lord Justice Geoffrey
Lane’s judgment in J Evans & Son (at 1085): ‘The express new term must be taken to override the printed
condition.’ Similarly, Lord Justice Scott in Couchman v Hill stated (at 559) ‘[T]here was clearly an oral offer
of a warranty which overrode the stultifying condition in the printed terms. . .’, and his concern about the
substantive control of standard terms becomes evident in the final paragraph of his judgment (‘The printed
condition that the vendor will take no responsibility for errors of description . . . is reasonable for visible
defects, but for qualities or attributes which are invisible it is not reasonable. It may well become a trap for the
unwary.’).
Such considerations do not equally apply to subsequent, supposedly ‘collateral’ agreements.
Contemporaneous oral agreements may well form part of a partially oral, partially written contract. Subsequent
oral agreements can only do one of two things: either they try to modify an earlier written agreement; or they
can leave the earlier written agreement intact and relate to a different, independent subject matter. If they relate
to the same subject matter, ie if they contradict the earlier written agreement, then they are necessarily attempts
at modification of the original contract, and calling them collateral contracts only attaches a new label in order
to circumvent the rules on modifications. Such sophistry should be avoided.
Additionally, for subsequent oral agreements, there is no need to muddy the distinction between
modifications and collateral contracts. The reasons that drove the courts to accept ‘contradictory’ collateral
contracts in the first place are the same ones that are addressed by waiver and estoppel (below 8); this
functional equivalence can be clearly seen in Couchman v Hill. Indeed, it has been convincingly suggested (by
Guest in Chitty on Contracts § 12-103 in fn 450) that in the cases of contemporaneous contradictory collateral
agreements ‘the same result could be reached by application of the principle of promissory estoppel . . ., but it
would appear that that principle may not extend to pre-contractual negotiations . . .’ This bar is irrelevant for
subsequent oral agreements. Therefore, there is no reason to interpret subsequent oral agreements that
contradict the original written contract as ‘collateral contracts’. Either the subsequent oral agreement deals
with a subject matter different from the one of the original contract, in which case there is a genuine collateral
agreement; or the agreement contradicts or qualifies the original contract. If the latter is the case, then it is a
variation, and if this variation lacks the required form, then the only question is whether the rules on waiver or
estoppel (below 8) apply; if they do not, then there is no justification for saving what is an invalid variation
under the name of a ‘collateral agreement’.
Moreover, the case City of Westminster Properties v Mudd [1959] Ch 129, which is often cited as a
‘contradictory collateral contract’ case, eg by Guest in Chitty on Contracts § 12-103, Peel Treitel Law of
Contract 6-026 and the Court in AS Klaveness, above n 172) arguably applied estoppel principles; Harman J.
did not mention the word ‘collateral’ at all (at least not in the report cited).
47
marginally more precise than the ‘sufficiently substantial’ criterion for the distinction
between modifications and novations.174 On the other hand, a statutory provision enforcing
NOM-clauses against novations is possible, because a novation requires a rescission of the
original contract that can by statute be subjected to the NOM-clause, whereas it would not
seem possible to prevent parties from informally concluding genuinely collateral
agreements.
While (alleged) novations and collateral contracts may occasionally undermine
NOM-clauses, they are relatively rarely invoked. A much greater challenge to the
effectiveness of NOM-clauses arises from waiver and estoppel or ‘good faith’ defences, to
which I now turn.

8. Estoppel, Waiver and Venire Contra Factum Proprium

Even where NOM-clauses are generally given effect, and where the ‘modification’ is not in
reality effective as a novation or collateral contract, there remains the question whether
purported oral modifications can be completely disregarded, or whether they may have some
residual effect.
One can illustrate the need for exceptionally attributing some residual effect to oral
modifications by comparing the NOM-clause (as a ‘private Statute of Frauds’) with the
historical Statute of Frauds 1677. While the Statute of Frauds addressed one mischief,
namely perjured evidence about non-existent agreements,175
it was capable of giving rise to another: that a party, making and acting on what was
thought to be a binding oral agreement, would find his commercial expectations
defeated when the time for enforcement came and the other party successfully relied
on the lack of a written memorandum or note of the agreement.176
Mutatis mutandis, there is a danger that one party believes an informal agreement to be
valid, where, because of the NOM-clause, it is not. In some circumstances, it would be
inequitable for the other party to go back on the statements made in the informal agreement.
To give an example:177 A and B agree in a contract containing a NOM-clause that A
is to deliver widgets to B in 10 instalments; A is to manufacture the widgets to the design

174
Above n 156.
175
Above, text accompanying nn 112, 113.
176
Actionstrength Limited v International Glass Engineering [2003] UKHL 17, [2003] 2 AC 541, 545 [2]
(Lord Bingham).
177
The following example is virtually identical with the standard examples, see eg CISG Secretariat
Commentary on the 1978 Draft, Article 27, Example 27A; also cf UNIDROIT Principles 2.1.18, illustration 2;
48
specification X. After the first instalment, B asks A to manufacture the widgets to the design
specification Y instead of X. A agrees, but the modification is not put into writing. A
delivers widgets according to the design specification Y in the second, third, fourth and fifth
instalment. B rejects the fifth instalment and demands damages because the original contract
specifically demands design specification X, and any oral modifications are invalid because
of the NOM-clause.
It seems obvious that holding A to the original contract under these circumstances
would be an objectionable result: after all, A has changed the design specification at B’s
explicit request. To avoid such injustice, all jurisdictions that enforce NOM-clauses provide
for certain exceptions to protect the party that believes the informal agreement to be binding.
These exceptions are usually called ‘waiver’ or ‘estoppel’. Unfortunately, both these terms
are notoriously ambiguous in legal discourse.178
UCC § 2-209(4) states: ‘Although an attempt at modification or rescission does not
satisfy the requirements of subsection (2) or (3) it may operate as a waiver.’179 The US
courts are divided on how to interpret ‘waiver’ in this provision. Frequently, ‘waiver’ is
defined as the ‘voluntary or intentional relinquishment of a known right’;180 alternatively, it
can be confined to the excuse of a condition.181 Under both definitions, the focus is on the
person relinquishing the right; there is no requirement for the other party to ‘rely’ on the
relinquishment. Nevertheless, the majority of US courts appears to require reliance for a
waiver under UCC § 2-209(4) to be effective;182 however, a ‘clear and unequivocal’ waiver

see also the illustration for DCFR II.-4:105(2) in von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol
I, 288–9 (for a discussion, see text before and after n 251).
178
cf, for England, Mardorf Peach & Co v Attica Sea Carriers (The Laconia) [1977] AC 850, 871: ‘[T]he
word “waiver,” like “estoppel,” covers a variety of situations different in their legal nature, and tends to be
indiscriminately used by the courts as a means of relieving parties from bargains or the consequences of
bargains which are thought to be harsh or deserving of relief’; Wilken, Waiver, Variation and Estoppel para
3.01 with numerous references; T Dugdale & D Yates, ‘Variation, Waiver and Estoppel – A Re-Appraisal’
(1976) 39 Modern Law Review 680 (distinguishing, at 681–3, six situations); GH Treitel in Chitty on
Contracts § 3-081 (who, because of the ambiguity of ‘waiver’, prefers the term ‘forbearance’). See generally
the detailed accounts in Wilken, ibid paras 3.01–16 and Snyder, ‘Concept of change’ [1999] Wisconsin Law
Review 607, 624–9 and passim. From the various definitions or classifications, we can exclude the category
‘waiver by election’ as irrelevant for present purposes. Also, we can disregard those older cases that used
‘waiver’ as synonymous with variation.
179
The 2003 revision of the UCC substituted the ‘may’ for ‘can’.
180
In the US, see eg, Rhorer v Raytheon Engineers and Constructors, 181 F.3d 634, 645 (5th Cir 1999); for
England, Wilken, Waiver, Variation and Estoppel para 3.15; E McKendrick, Goode on Commercial Law (4th
edn, London: Penguin Books, 2010) 117.
181
Farnsworth, Contracts, § 7.6 (text accompanying fn 19 and 20); Snyder, ‘Concept of change’ [1999]
Wisconsin Law Review 607, 625–6.
182
The best-known case is Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280 (7th Cir 1986). In
this case, Judge Richard Posner held for the majority that reliance was a necessary ingredient for a ‘waiver’
under UCC § 2-209(4), while Judge Easterbrook in a dissenting opinion vigorously argued for interpreting
‘waiver’ in the traditional way as ‘a known relinquishment of a right’. A distinguished commentator remarks
49
may suffice even in the absence of reliance.183 The view requiring reliance essentially
confounds the common law doctrine of waiver with the equitable doctrine of ‘estoppel’, a
doctrine which requires not only one party to make a representation or promise, but also
some form of reliance by the other party, before the first party is estopped from going back
on the statement. In the context of UCC § 2-209, however, the two views are not often far
apart. UCC § 2-209(5) allows the retraction of the waiver for an executory part of the
contract by reasonable notification, but makes the waiver irretractable as soon as there has
been a material change of position in reliance on the waiver.184 Even if one does not require
reliance for a waiver in UCC § 2-209(4), this waiver will as such only preclude the party
from insisting on the knowingly relinquished right for executed portions of the contract.
Where there was no reliance, the relinquishing party is free to retract the waiver for
executory portions of the contract. In contrast, where there was reliance, the waiver may be
irretractable under UCC § 2-209(5).
Californian Law allows recourse both to waiver and to estoppel.185 The CISG186 and
the UNIDROIT Principles187 provide for an exception explicitly only for cases in which
there was reliance. In Germany, the enforceability of the double NOM-clause is similarly

on the influence of Wisconsin Knife: ‘Where . . . there has been no substantial change in position in reliance on
the oral modification, the courts are split as to whether a waiver will be found, with a probable majority
following an influential, though probably incorrect, Seventh Circuit decision.’ (RA Lord in Williston on
Contracts Vol 10 (4th edn, West 1993 & Supp 2010) § 29:42). For a Federal Court of Appeals decision
following the minority view in Wisconsin Knife, see eg BMC Industries v Barth Industries, 160 F.3d 1322,
1333-4 (11th Cir. 1998). See also Farnsworth, Contracts, § 7.6 (text accompanying fn 19 and 20).
183
Cloud Corp v Hasbro, 314 F.3d 289, 297–8 (7th Cir 2002).
184
If NOM-clauses are unenforceable in England (under World Online, above 2), then the question of waiver
or estoppel does not arise in this jurisdiction. If they are (under Asif, above 3), the situation would presumably
be similar to the US American position: Waiver is in principle revocable, but may become irrevocable once
there has been reliance; and equitable estoppel requires reliance from the start, and is also generally suspensive
only. See, generally, GH Treitel in Chitty on Contracts § 3-082-3 and § 3-085-104 (esp § 3-104 on the
similarity between waiver and equitable estoppel).
185
For the text of § 1698(d) of the Californian Civil Code, see n 172.
186
CISG Art 29(2), second sentence, provides: ‘However, a party may be precluded by his conduct from
asserting such a provision to the extent that the other party has relied on that conduct.’ The exception is ‘based
on principles contained in the . . . Mißbrauchseinwand of German law, the “nemo suum venire contra factum
proprium” principle of Roman law, or the doctrine of waiver and estoppel of Anglo-American law’, S Eiselen,
‘Modification’ 165 comment h.
187
UNIDROIT Principles Art 2.1.18, second sentence: ‘However, a party may be precluded by its conduct
from asserting such a clause to the extent that the other party has reasonably acted in reliance on that conduct.’
As under German law (see n 188), the exception is considered to be an application of the general principle
prohibiting inconsistent behaviour, enshrined in UNIDROIT Principles Art 1.8, see Comment to Art 2.1.18; J
Kleinheisterkamp in S Vogenauer & J Kleinheisterkamp (eds), Commentary on the UNIDROIT Principles of
International Commercial Contracts (PICC) (Oxford: Oxford University Press 2009) Art 2.1.18 para 2
(‘exception of venire contra factum proprium’).
50
subject to an exception, derived from – the English lawyer will add: horribile dictu – ‘good
faith’.188
Of course, one should not apply any of these exceptions too liberally, lest one give
with one hand the enforceability of NOM-clauses, only to take it away with the other.189
First, not each and every purported modification should be considered sufficient to
invoke the exception. 190 The majority of US courts require reliance or a ‘clear and
unequivocal’ waiver for UCC § 2-209(4). Even if ‘waiver’ in UCC § 2-209(4) should not
require reliance, this subsection only states that an attempt to modify or rescind may (until
2003: can) operate as a waiver, not that every such attempt does operate as a waiver.191
For CISG Article 29(2) and UNIDROIT Principles 2.1.18, there must at least be
reasonable reliance by one party on statements or conduct by the party that seeks to enforce
the NOM-clause.192
While it is easy to state that the standard of ‘reasonable reliance’ applies, it is much
more difficult to specify in which circumstances it is reasonable to rely on informal
statements or conduct, 193 given that the contract contains a NOM-clause. A perfectly

188
See below text accompanying n 195. For the avoidance of doubt, it should be pointed out that good faith in
German law is not as amorphous a concept as is sometimes supposed. Roughly comparable to the system of
equity, there are various well-defined subcategories of the good faith principle established by case law. The
applicable subcategory here is the prohibition of ‘venire contra factum proprium’ – the prohibition against
self-contradictory behaviour. Indeed, the good-faith exception to double NOM-clauses in German law appears
to be markedly narrower than the Common Law counterpart of waiver and estoppel, see below text
accompanying n 195.
189
Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280, 1286 (7th Cir 1986); Jolls, ‘Contracts as
bilateral commitments’, (1997) 26 Journal of Legal Studies 203, 206, 230–1; Boergen, ‘Effektivität’ [1971]
Betriebs-Berater 202, 204; Kleinheisterkamp in Vogenauer & Kleinheisterkamp PICC Art 2.1.18 para 2: ‘The
proviso . . . is problematic since it partially reintroduces the uncertainty that the parties had intended to
eliminate. This uncertainty must be reduced as far as possible by a restrictive interpretation . . . Cases in which
this is to be affirmed should be extremely rare.’
190
Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280, 1286 (7th Cir 1986).
191
Cloud Corp v Hasbro, 314 F.3d 289, 297–8 (7th Cir 2002). As a limiting criterion, Judge Easterbrook
suggested in Wisconsin Knife: ‘It might be sensible to treat claims of oral waiver with suspicion and insist on
waiver by course of performance . . . Waiver implied from performance is less prone to manipulation.’ (781
F.2d at 1292).
192
In contrast to UNIDROIT Principles Art 2.1.18., which explicitly requires ‘reasonable reliance’, CISG Art
29 requires ‘reliance’ without the qualification ‘reasonable’. Nevertheless, the predominant – and preferable –
view is that reliance under CISG Art 29 must be reasonable as well; the standard of reasonableness can
arguably be supplied as a general principle under CISG Art 7(2). S Eiselen, Guide to Article 29 – Comparison
with PECL, <http://cisgw3.law.pace.edu/cisg/text/peclcomp29.html> (2002), comment j; idem, ‘Modification’
166 comment i (although this author grounds the requirement of reasonableness on good faith, CISG Art 7(1)).
Of course, one could also argue e contrario that elsewhere the CISG drafters qualified reliance by adding a
reasonableness-criterion (in Art 16(2)(b) and 35(2)(b)), so that the omission in Art 29(2) may have been
deliberate; but this appears unlikely. The question is raised, but not answered in Hillman, ‘Article 29(2)’, 21
Cornell International Law Journal 449, 460 (1988).
193
Both CISG Art 29(2) and UNIDROIT Principles Art 2.1.18 mention only ‘conduct’, not ‘statements’. It has
been suggested, however, that reliance on statements may also qualify, del Pilar Perales Viscasillas,
‘Modification’, 25 Journal of Law and Commerce 167, 178 (2005). On the one hand, it is true that it would be
curious if parties could imply the other party’s consent from their conduct, but could not rely on explicit
51
rational actor would always bear the existence of the NOM-clause in mind and would
therefore never rely on any informal agreements.194 This shows that the reliance-based
exceptions envisage less than the standard of a perfectly rational actor; otherwise, the
exception would never apply. On the other hand, if reliance on all informal agreements were
sufficient, then the reliance-based exception would swallow the rule of the enforceability of
NOM-clauses. There does not appear to be a bright-line test to isolate those cases in which
reliance is reasonable. Instead, it seems that one has to resort to the somewhat
unsatisfactorily vague standard that going back on the informal statement or conduct must
be inequitable. While it would be more satisfactory if one could draw a bright line, this is
hardly the only area of law in which a vague standard is necessary to do justice to the
vagaries of life.
The German Federal Court of Justice opted for a very narrow exception in the 1976
decision, which established the enforceability of double-NOM clauses:
[W]here the form requirement was agreed between the parties, the good-faith
defence only applies if extraordinary circumstances justify this. […] Where, as here,
merchants want to exclude the relevance of oral statements and conduct in an
individually negotiated agreement, evidently knowing the case law [scil.: that
‘simple’ NOM clauses are ineffective], the good-faith defence can generally only be
invoked where the other party has deliberately prevented compliance with the form
requirement.195
While it does not seem possible to confine the reliance-based exceptions exclusively to the
one case that the party invoking the NOM-clause has deliberately prevented compliance
with the agreed form requirement,196 it is a good example for the exceptional situation in
which there is more than a mere informal agreement to make reliance appear reasonable.

statements about such consent; on the other hand, see Judge Easterbrook’s suggestion (n 191), which can claim
the support of the adage ‘actions speak louder than words’. Also cf Atiyah, Rise and Fall, 205, stating in a
different context: ‘Even in modern times, it is a great deal easier to prove an agreement where the plaintiff
has . . . relied upon a promise by changing his position in some way. Acts of this kind are usually objective
facts, provable by disinterested witnesses . . . But a bare agreement, unless it is evidenced in writing, is clearly
more difficult to prove . . .’
194
cf Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 460 (1988).
195
Bundesgerichtshof 2 June 1976, case VIII ZR 97/74, 66 Entscheidungen des Bundesgerichtshofs in
Zivilsachen 378, 383, [1976] Neue Juristische Wochenschrift 1395 (author’s translation, emphasis added);
followed by OLG Koblenz 8 November 2004, case 12 U 244/03 [2005] BeckRS 05347 (part II.2.c).
196
The Court left a backdoor open by inserting the word ‘generally’. But see, for the interpretation of Art
2.1.18 of the UNIDROIT Principles: Kleinheisterkamp in Vogenauer & Kleinheisterkamp PICC Art 29 para 2:
‘The exception can only operate where it is clear that one party induced the other party to believe in a waiver
of the form requirement and was conscious that the other party would act to its detriment in reliance thereon.’
52
 Secondly, the consequences of a waiver or estoppel will usually not be as far
reaching as those of a contract modification would be. A waiver or estoppel generally
suspends the strict obligations under the (unmodified) contract, and does not change the
obligations on a lasting basis.
For example, UCC § 2-209(5) provides:
A party who has made a waiver affecting an executory portion of the contract
may retract the waiver by reasonable notification received by the other party
that strict performance will be required of any term waived, unless the
retraction would be unjust in view of a material change of position in reliance
on the waiver.
Similarly, the CISG- and UNIDROIT-provisions make clear that the party invoking the
NOM-clause is only precluded from doing so ‘to the extent’ that the other party has
relied.197 Thus, in the example above, B could demand for future deliveries a reversion to
the specifications on which the parties had originally agreed (X), unless A had, for example,
relied by making substantial investments into changing the manufacturing process to
specification Y.
In commercial practice, parties often want to avoid the NOM-clause from being
undermined by these exceptions. They try to achieve this result by adding a ‘no-oral-waiver’
clause to the NOM-clause.198 The effect of such no-oral-waiver clauses is even less certain
than that of NOM-clauses. At a minimum, a no-oral-waiver clause will make courts or
arbitrators think twice whether it was reasonable to rely on an informal agreement to
modify a contract. Some US American courts have even enforced no-oral-waiver clauses by
refusing to consider waiver irrespective of the conduct of the parties. 199 Similarly, an
English judge stated that ‘it appears that there is no general principle of law that one cannot
restrict the operation of the doctrine of waiver by contract.’200 Other US American courts,

197
See also the CISG Secretariat Commentary on the 1978-Draft, Art 27 (the Draft-equivalent of CISG Art 29),
para 9: ‘It should be noted that the party who wishes to assert the provision in the contract which requires any
modification or abrogation [now: termination] to be in writing is precluded from doing so only to the extent
that the other party has relied on the conduct of the first party. This may mean in a given case that the terms of
the original contract may be reinstated once the first party denies the validity of the non-written modification.’
(emphasis added).
198
Snyder, ‘Concept of Change’ [1999] Wisconsin Law Review 607, 638–70.
199
For decisions enforcing the no-oral-waiver clause, see eg South Hampton v Stinnes Corp, 733 F.2d 1108,
1117 (5th Cir 1984); National Data Payment Systems v Meridian Bank, 212 F.3d 849, 855 (3rd Cir 2000);
Techne Corp and Research and Diagnostic Systems v Amgen, 2001 WL 1690062 (D.Minn.2001) (slip copy at
5).
200
State Securities v Initial Industry [2004] All ER (D) 317 (Jan) [56]. However, one should note, first, that
this concerned a waiver by election, and secondly, the limited scope of the clause considered: ‘I can . . . see no
reason in principle why the parties to . . . [a] commercial contract, should not be free to stipulate that a
53
however, approach no-oral-waiver clauses with more suspicion and allow these clauses to
be waived in appropriate cases;201 similarly, Professor McKendrick has noted that ‘it cannot
be assumed that the courts in all cases will give effect to a term of the contract which
purports to exclude or limit the operation of the doctrine of waiver’.202 Under the CISG, the
question whether the parties can derogate from the second sentence of Article 29(2) is
contentious; while Article 6 generally permits derogations from the CISG or parts thereof,
some argue that the second sentence of Article 29(2) is, as an emanation of the prohibition
against self-contradictory behaviour, ius cogens.203
Under German law, the position is clear: the venire contra factum proprium-defence
of good faith, while narrow in scope, is ius cogens and thus cannot be excluded by the
parties.204
Do the exceptions based on waiver or estoppel swallow the rule of enforceability of
the NOM clause? The possibility of relying on oral statements to prove the exception
certainly reduces the effectiveness of the NOM clause.205 Where there is a possibility that
the exception can be invoked, the parties may have to go to trial even in jurisdictions where

particular act, such as payment of a rental instalment should not be taken to waive a right to terminate for an
earlier breach.’ Ibid [57]. Excluding particular conduct from being treated as a waiver is for obvious reasons
much less problematic than a wholesale no-oral-waiver clause.
201
eg Christian Dior-New York v Koret, 792 F.2d 34, 39–40 (2nd Cir. 1986), distinguished by RBFC One v
Zeeks, 367 F.Supp.2d 604, 611–2 (SDNY 2005) (where the clause was formulated differently and the intention
to waive was unclear); for further references, both for the enforcement and lack of enforcement see CL Kunz,
‘Teaching First-Year Contracts Students How to Read and Edit Contract Clauses’ 34 University of Toledo
Law Review 705, 710 (2003).
202
McKendrick in Chitty on Contracts § 22-045. Of course the question of no-oral-waiver clauses does not
arise in our context if NOM-clauses are not enforced in the first place, as may be the case in England under
World Online v I-Way [2002] EWCA Civ 413 (but see above 3 at nn 72–79). For Scots law, see WW McBryde,
The Law of Contract in Scotland (3rd edn, Edinburgh: Thomson/W. Green, 2007) § 25-07. Also cf F de Ly,
‘Commercial law as a refuge from contract law: A comparative and uniform law perspective’, 45 Wayne Law
Review 1825, 1862 (2000) (noting that ‘contracts often contain anti-waiver provisions to provide additional
protection. Any such protection is, however, not absolute, but will, to a large extent, depend on the factual
elements involved in any such conduct.’)
203
For the possibility of derogatation from Art 29(2) (second sentence): Schlechtriem in Schlechtriem &
Schwenzer, Kommentar UN-Kaufrecht, Art 29 para 10a. Against: Hillman, ‘Article 29(2)’, 21 Cornell
International Law Journal 449, 461–2 (1988) (arguing that otherwise every NOM-clause would implicitly
derogate from the reliance-provision; while this argument could be countered by requiring an explicit no-oral-
waiver clause, this would in turn require an explanation why here an explicit derogation should be necessary
when Art 6 usually allows implicit derogations); Geldsetzer, Einvernehmliche Änderung, 156; U Magnus in
Staudingers Kommentar (2005) Art 29 CISG para 21.
204
Larenz & Wolf, Allgemeiner Teil, § 34 para 49; D Looschelders & D Olzen in Staudingers Kommentar
(2005) § 242 paras 107–9; Grüneberg in Palandt § 242 para 6 (noting, however, that the parties may exclude
certain, specifically defined conduct from being considered an infringement of the good faith principle; this
may narrow the gap to the English position, cf n 200).
205
cf Judge Posner’s remarks in Autotrol Corp v Continental Water Systems Corp, 918 F.2d 689, 692 (7th Cir
1990).
54
the NOM-clause is given effect; the court will usually not grant a motion to dismiss or an
application for a summary judgment.206
Even though the reliance-based exceptions limit the effectiveness of NOM-clauses,
it still seems preferable to give effect to NOM-clauses and add a narrow exception for
waiver or estoppel cases (or good faith), rather than to give no effect to NOM-clauses at all.
Courts are sufficiently aware of the problem that the exception should not swallow the rule,
and will be careful to confine the exception within narrow bounds by applying both a strict
test to the requirements of reasonable reliance, and by limiting the consequences of the
exception to the extent required.

9. NOM-clauses in Standard Terms and Conditions; and the Distinction Between

B2B, B2C, and C2C
,p>Several jurisdictions that enforce individually negotiated NOM-clauses provide for
special rules for NOM-clauses in non-negotiated standard terms and conditions, at least if
they are used in relationships other than between businesses.

United States
In the US, UCC § 2-209(2) states: ‘A signed agreement which excludes modification or
rescission except by a signed writing cannot be otherwise modified or rescinded, but except
as between merchants such a requirement on a form supplied by the merchant must be
separately signed by the other party.’

Germany
In Germany, a NOM-clause contained in standard terms and conditions has to pass two
separate, but not entirely independent hurdles, and most NOM-clauses fail one or both: (i)
even if the NOM-clause is incorporated into the contract as one of several standard terms
and conditions, and is valid under the fairness test (see below (ii)), individually negotiated
agreements still take precedence over conflicting terms and conditions pursuant to § 305b of
the German Civil Code; (ii) standard terms and conditions have to pass the fairness test
under §§ 307-9 of the German Civil Code.

206
This possibility of estoppel was an alternative ground on which Sedley LJ held that the World Online case
should go to trial, see [2002] EWCA Civ 413 [13]. In Asif, on the other hand, summary judgment was granted,
because the estoppel-based defence was considered to be fanciful. See United Bank v Asif 2000 WL 456 [24]–
[25] (‘I cannot think that any person, no matter what their degree of gullibility, would say, “Ah yes, that
sounds like a plausible story” ’).
55
 Even where NOM-clauses in standard terms and conditions are valid, they remain
generally ineffective in Germany because any subsequent individually negotiated
modifications will take precedence over all standard terms and conditions, § 305b of the
German Civil Code.207
As to the second hurdle, the application of the fairness test is not straightforward.
§ 309 no 13 of the German Civil Code provides that ‘a clause stipulating for a stricter form
than writing’ is invalid in standard terms and conditions. An argumentum e contrario
suggests that clauses stipulating only for the written form – as practically all NOM-clauses
do – are therefore not in themselves unfair. And this is indeed the starting point for the
courts: clauses stipulating for written form are not in themselves unfair; instead, the fairness
depends on the formulation and scope of the specific clause.208 However, as soon as the
clause gives the impression that all agreements not complying with the form would be
invalid, they are misleading – because, as described above, 209 individually negotiated
agreements will take precedence over standard terms and conditions. This misleading
impression makes most NOM-clauses in standard terms and conditions ‘unfair’, because
they could prevent the party against whom these terms are used from enforcing an informal,
but individually negotiated agreement.210
One can draft a NOM-clause that avoids giving this misleading impression – but it
would hardly ever be worth the effort: Such a NOM-clause would be valid under the
fairness test, but subsequent informal, yet individually negotiated, agreements would usually

207
Bundesgerichtshof 21 September 2005, case XII ZR 312/02 [2006] Neue Juristische Wochenschrift 138–9;
Bundesarbeitsgericht 20 May 2008, case 9 AZR 382/07 [2009] Neue Juristische Wochenschrift 316, 318
(para. 29); Oberlandesgericht Düsseldorf 1 June 2006, 10 U 1/06 [2006] BeckRS 07261; see also
Bundesgerichtshof 15 May 1991, VIII ZR 38/90 [1991] Neue Juristische Wochenschrift 1750, 1751. For
similar rules see PECL Art 5:104, Acquis Principles Art 6:202, and DCFR II.-8:104, all providing that ‘[t]erms
which have been individually negotiated take preference over those which have not.’
208
Bundesgerichtshof 7 October 1981, case VIII ZR 229/80 [1982] Neue Juristische Wochenschrift 331, 333
(stressing the advantages of NOM-clauses for certainty and evidentiary proof); Bundesgerichtshof 31 October
1984, case VIII ZR 226/83 [1985] Neue Juristische Wochenschrift 320, 322; Bundesgerichtshof 26 March
1986, case VIII ZR 85/85 [1986] Neue Juristische Wochenschrift 1809, 1810; Bundesgerichtshof 15 May 1991,
case VIII ZR 38/90 [1991] Neue Juristische Wochenschrift 1750, 1751–2; Bundesgerichtshof 15 February
1995, case VIII ZR 93/94 [1995] Neue Juristische Wochenschrift 1488, 1489; Bundesgerichtshof 27
September 2000, case VIII ZR 155/99 [2001] Neue Juristische Wochenschrift 292; Bundesgerichtshof 10 May
2007, case VII ZR 288/05 [2007] Neue Juristische Wochenschrift 3712, 3713 [19].
209
n 207.
210
Bundesgerichtshof 31 October 1984, case VIII ZR 226/83 [1985] Neue Juristische Wochenschrift 320, 322;
Bundesgerichtshof 15 May 1991, VIII ZR 38/90 [1991] Neue Juristische Wochenschrift 1750, 1751–2;
Bundesgerichtshof 15 February 1995, case VIII ZR 93/94 [1995] Neue Juristische Wochenschrift 1488, 1489;
Bundesgerichtshof 27 September 2000, case VIII ZR 155/99 [2001] Neue Juristische Wochenschrift 292–3;
Oberlandesgericht Rostock 19 May 2009, case 3 U 16/09 [2009] Neue Juristische Wochenschrift 3376; see
also Bundesgerichtshof 10 May 2007, case VII ZR 288/05 [2007] Neue Juristische Wochenschrift 3712, 3713
[19] (obiter).
56
take precedence over the NOM-clause pursuant to § 305b of the German Civil Code.211
Only in the area of employment law, which has a separate doctrine for forming agreements
without individual negotiations by course of conduct,212 do appropriately narrowly drafted
NOM-clauses in standard terms and conditions play a significant role.213

CISG
During the deliberations of Article 27 of the Draft for the CISG (now CISG Article 29),
Italy had proposed to insert a third paragraph:
(3) The preceding paragraph shall not apply where the provision requiring
modifications or abrogations of the contract to be in writing is contained in general
conditions prepared by one party and that party either directly or through an
authorized agent orally agrees to modify or abrogate his general conditions.214

Professor Michael Joachim Bonnell, who submitted the proposal, explained that:
[a]lthough that principle [now contained in CISG Article 29(2)] was unknown in
Italian law, under which a contract could be derogated from by oral agreement as
well, his delegation was nevertheless prepared to accept it, on the grounds that it
offered a sound solution for other legal systems and for international trade in general.
However, his delegation considered that it was necessary to limit the principle to [sic]
a particular situation which often occurred in trade practice, when the requirement of

211
There is one case that does not fit the analysis described in the text. In 2007, the Federal Court of Justice
upheld a NOM-clause as not being unfair, and rejected the view that the individually negotiated oral
agreement to modify the contract should take precedence over the NOM-clause included in the terms and
conditions (Bundesgerichtshof 10 May 2007, case VII ZR 288/05 [2007] Neue Juristische Wochenschrift 3712,
3713). While the Court’s reasons for not considering the clause unfair may be persuasive, it remains unclear
why the individually negotiated oral agreement on the contract modification should not take precedence over
the standard terms and conditions. Friends of public choice theory might point to the fact that the user of the
standard terms and conditions in the case was the Federal Republic of Germany. The explanation seems to be
that the Federal Court of Justice in this case appears to require a separate agreement to disapply the NOM-
clause (cf para 26); while some courts have required such a separate agreement (above nn 49, 50), the majority
of courts have not (above n 42), and where § 305b of the Civil Code applies, it would be highly unusual to let
the standard terms trump a conflicting individually agreed modification (see Bundesgerichtshof 21 September
2005, case XII ZR 312/02 [2006] Neue Juristische Wochenschrift 138–9; Bundesarbeitsgericht 20 May 2008,
case 9 AZR 382/07 [2009] Neue Juristische Wochenschrift 316, 318 [29]).
212
So-called ‘betriebliche Übung’. A minority view considers the ‘betriebliche Übung’ as a standard formation
of contract by offer and acceptance by conduct; this minority view would presumably apply § 305b of the
German Civil Code here as well.
213
cf Bundesarbeitsgericht 24 June 2003, 9 AZR 302/02 [2003] Neue Juristische Wochenschrift 3725, 3727;
Bundesarbeitsgericht 20 May 2008, case 9 AZR 382/07 [2009] Neue Juristische Wochenschrift 316, 317; see
also S Lingemann & M Gotham, ‘Doppelte Schriftformklausel – gar nicht einfach’ [2009] Neue Juristische
Wochenschrift 268; M Bloching & D Ortolf, ‘Schriftformklauseln in der Rechtsprechung von BGH und BAG’
[2009] Neue Juristische Wochenschrift 3393.
214
See document A/CONF.97/C.1/L.68.
57
 written abrogation or modification had not been specifically accepted by the parties
to a contract, but had merely been included in the general conditions drawn up
unilaterally by one of them but nevertheless forming [sic] part of the contract. It
often happened that the same party agreed orally to certain modifications in [sic] the
conditions. In that case the oral agreement modifying the contents of the general
conditions should prevail, and the general principle in paragraph 2 should not apply.
Furthermore, as the party who had drawn up the general conditions was sometimes
represented by an authorized agent in the actual negotiations, oral modifications
agreed to by the agent might be repudiated by his principal in the event of a dispute.
The paragraph he had proposed was intended to prevent the economically weaker
party from falling into that kind of trap.215
The proposal was rejected, after the Bulgarian delegate had stated that ‘it appeared to be
more suited to consumer contracts than to international commercial contracts. The parties to
commercial contracts usually [enjoy] equal bargaining power and [do] not need the defence
offered by the amendment’.216
Later, Sweden and Denmark unsuccessfully tried to avoid being bound by CISG
Article 29(2), because of the danger of abuse by hiding away NOM-clauses in standard
terms and conditions.217

B2B, B2C, AND C2C

A closely related question is whether one should enforce NOM-clauses in Business-to-
Business (B2B) relationships, Business-to-Consumer (B2C) relationships,, and Consumer-
to-Consumer (C2C) relationships, without differentiation. The policy reasons for enforcing
NOM-clauses are certainly stronger for agreements in B2B relationships than for contracts
between a business and a consumer. First, the need for NOM-clauses will usually be
stronger in B2B relationships, in particular to ensure certainty in high-stake, long-term

215
Summary records of meetings of the First Committee, 13th meeting on 19 March 1980, para 56. The Italian
legal position with regard to individually negotiated NOM clauses appears to have changed in the meantime in
favour of their enforcement (see text accompanying n 239).
216
Ibid, para 58. Another reason was that the CISG had carefully avoided the controversial issue of standard
terms and conditions, and the proposed amendment was seen to reopen this issue. Ibid, para 59 (comment by
Czechoslovakian delegate) and P Schlechtriem, Uniform Sales Law – The UN Convention on Contracts for the
International Sale of Goods (Vienna: Manz 1986) 45 in fn 147.
217
See Document A/CONF.97/L.9, 7 April 1980: ‘Paragraphs (1) and (2) should be voted upon separately
(rule 38). Paragraph (2) (= agreed written form) should be either moved to Part II of the Convention (as article
22 bis) or (if not moved) not to [sic] be adopted. Note: . . . paragraph (2) [deals] with written form as agreed
between parties for modifications of a contract. Provisions of the latter type – often appearing as small print in
General Conditions issued by the stronger party – may easily be abused and should not be encouraged.’
58
relationships. Secondly, the danger of catching the other party unawares will be larger in a
B2C relationship.218

Comparative Remarks on NOM-clauses in Standard Terms

The details are not of paramount importance here. From a comparative perspective, one
should note that several jurisdictions that enforce NOM-clauses treat NOM-clauses in
standard terms and conditions differently, at least if used by businesses against consumers.
The rationale is intuitive: in most cases, in which NOM-clauses are litigated, the parties
were simply not aware of the NOM-clause at the time they modified their contract and
believed the modification to be effective. The danger that the parties overlook or forget the
existence of the NOM-clause is particularly great if it is hidden away in the small print. And
while one may demand that businesses be aware of the possibility of NOM-clauses in
standard terms, given the frequent use of such clauses in commercial practice, consumers
will usually not think of the possibility of such a clause, and may therefore require special
protection.
It could be that part of the explanation for the hostility with which NOM-clauses are
treated in some jurisdictions219 is this suspicion against surprising NOM-clauses in standard
terms.220 If this should be the case, then the right response for a legislator would be to
address NOM-clauses in standard terms differently, rather than extend this suspicion to
individually negotiated NOM-clauses, especially where the parties agreeing on the NOM-
clause are businesses well versed in commercial practice.221

10. DCFR II.-4:105

Let us now turn to the solution the drafters of the Draft Common Frame of Reference have
found for NOM-clauses. DCFR II.-4:105 provides for ‘Modification in Certain Form Only’
clauses:

218
See the quote above, text accompanying n 136.
219
Above 2.
220
cf Wisconsin Knife Works v National Metal Crafters, 781 F.2d 1280, 1286 (7th Cir. 1986) (Judge Richard
Posner): The common law position of not enforcing NOM-clauses ‘may have reflected a fear that such clauses,
buried in the fine print of form contracts, were traps for the unwary . . .’
221
Again, see the quote above at n 124. See also CL Knapp, ‘Opting Out or Copping Out? An Argument for
Strict Scrutiny of Individual Contracts’ 40 Loyola of Los Angeles Law Review 95, 101 (2006): ‘Dominance of
the document [scil.: achieved by merger and NOM-clauses] is not necessarily problematic in terms of the
parties’ intentions; indeed, it may be what both parties want . . . because each hopes to avoid costly arguments
later about the terms of their agreement. But that assumes that two parties, each with substantial bargaining
power, are actively negotiating terms. When the agreement is embodied in a standardized form, created by one
party and acceeded to by the other, different considerations come into play.’
59
 (1) A term in a contract requiring any agreement to modify its terms, or to terminate the
relationship resulting from it, to be in a certain form establishes only a presumption that any such
agreement is not intended to be legally binding unless it is in that form.
(2) A party may by statements or conduct be precluded from asserting such a term to the extent
that the other party has reasonably relied on such statements or conduct.

DCFR II.-4:105(1): A Very Rebuttable Presumption

The first paragraph declares that a NOM-clause establishes ‘only a presumption’ that an
informal agreement is not legally binding. The comments (and the definitions222) clarify that
this is a ‘rebuttable’ presumption. So far it sounds as if the DCFR were willing to give effect
to NOM-clauses, unless the presumption was rebutted.
This first impression would, however, be misleading. The relevant question is what
is required to rebut the presumption. The comments explain:
It would be contrary to good faith to let the parties’ agreement to use a particular
form bind them to that form when later they have clearly made up their minds to use
another form. If, therefore, it can be shown that both parties agreed to a
modification of the contract terms or a termination of the contractual relationship,
but did not use the specified form, effect must be given to their agreement. This
applies even if in an individually negotiated clause in their contract they provided
that they would not give effect to an oral agreement to disregard the ‘no oral
modification’ clause.223
In other words, the presumption for the effectiveness of the NOM clause will be rebutted as
soon as there is proof for an oral agreement to modify the contract or to terminate the
contractual relationship. Or more concisely: the NOM clause does not have any effect at all.
For it goes without saying that a party relying on a contractual modification has, at the very
least, to prove that this modification was agreed on – this is true with or without a NOM-
clause.224 If this minimum requirement of proving an agreement on a modification is at the
same time sufficient to ‘rebut’ the ‘presumption’ that the modification is not intended to be
legally binding, then the NOM-clause is not worth the paper it is written on. DCFR II.-

222
See von Bar, Clive, Schulte-Nölke et al, DCFR Outline Edition, 562; idem, DCFR Full Edition, Vol I, 77.
223
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 288 (emphasis added). Similarly, the
Principes Contractuels Communs, 232-3, state: ‘La présomption étant simple, il incombe à la partie intéressée
d’établir que les cocontractants ont, d’un commun accord, décidé de s’affranchir de la clause en question (dite
“clause d’interdiction de modification orale”) pour décider, sans écrit, de la modification, voire de la résiliation,
du contrat.’
224
See above in and text accompanying nn 46, 47.
60
4:105(1) is an unnecessarily complicated way of describing an empty set. If the drafters did
not want NOM-clauses to have a binding effect, they could have said so.
In order to avoid a misunderstanding, one should note that the presumption in
DCFR II.-4:105 is a different kind of ‘presumption’ from the one known in many European
legal systems, such as § 125 of the German Civil Code, Article 1352 of the Italian Civil
Code, and Article 159(2) of the Greek Civil Code. These code provisions establish the
evidential presumption that an agreed form requirement goes ad substantiam rather than
only ad probationem; in other words, in case of doubt it is presumed that the parties when
agreeing on a NOM-clause want future informal juridical acts – namely informal
modifications – to be null and void. This presumption cannot – in contrast to the
presumption in the DCFR II.-4:105 – be rebutted by proving that the parties subsequently
agreed informally on a contract modification. Instead, the evidential presumption would
only be rebutted if it were shown that the parties at the time of agreeing on the NOM-clause
did not intend informal modifications to be null and void. The NOM-clauses discussed in
this paper, which state that ‘no modification . . . shall be validly made unless in writing . . .’,
are always meant to go ad substantiam; the evidential presumption could not be rebutted. It
is a separate issue whether a subsequent oral agreement then supersedes the NOM-clause –
which may be the case even if the latter was intended to be ad substantiam. The DCFR
confounds these separate issues. The evidential presumption in the national civil codes
strengthens the case for enforcing NOM-rules; the presumption contained in the DCFR
seeks to emasculate them.

Does Comparative Law Support the Rule in DCFR II.-4:105(1)?

We have seen that the DCFR does not stand alone in giving no effect to NOM-clauses:
After all, under the common law of most states in the US, and possibly in England, NOM-
clauses are unenforceable, and in Germany ‘simple’ NOM-clauses are unenforceable despite
a Code provision that looks vaguely similar to DCFR II.-4:105.225
Nevertheless, the decision of the DCFR-drafters is unexpected. Generally, they have
looked to harmonized contract and sales law, in particular the CISG and the UNIDROIT
Principles, for inspiration. Both the CISG and the UNIDROIT Principles give full effect to
NOM-clauses.226

225
Above 2.
226
The divergence from the CISG and UNIDROIT Principles is acknowledged in the comparative notes, von
Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 290, Note no 12. On the generally ‘strong impact’
61
 Given the deliberate departure from the provisions in CISG 29(2) and UNIDROIT
Principles 2.1.18, one would expect an elaborate explanation why the solution found in
these harmonized instruments was rejected in the DCFR. After all, the drafters of the CISG
and UNIDROIT Principles already extensively discussed the legal position in various legal
systems as well as the intrinsic merits of NOM-clauses. Also, there is a widespread
recognition that there is a commercial need for NOM-clauses – indeed, the comments to
DCFR II.-4:105 themselves remark in the very first sentence that NOM-clauses ‘often occur,
especially in long-term contracts’.227 Rendering NOM-clauses unenforceable leaves this
demand unfulfilled and defeats the expectations of the parties using them; one would hope
to see an explanation why this should be the best solution.
The reader looking for a discussion of the costs and benefits of enforcing NOM-
clauses will be disappointed. The passage quoted above is all the interested reader can find
on the motives of the drafters of the DCFR for rejecting the enforceability of NOM-
clauses.228 The only ‘reason’ advanced is that it would be ‘contrary to good faith’ to bind the
parties to their own agreement.229
If it was not a weighing of the respective costs and benefits of NOM-clauses, one
might suspect that a thorough comparative review led the drafters to the conclusion that
NOM-clauses should not be enforced. Indeed, the comparative notes start confidently with
the assertion that ‘[t]he rule in paragraph (1) which only gives evidential weight to written
modifications clauses is in accordance with the laws of most of the countries of the Union as
far as contracts in general are concerned.’230 This would probably be true if the ‘evidential
weight’ concerned the question whether an agreed form requirement was meant to go ad
substantiam or ad probationem; but, again, this is not what the DCFR drafters seek to
establish. They want to argue that in the majority of EU Member States NOM-clauses are
worthless, because they will be superseded by any subsequent informal agreement.

of the CISG on the evolving European Contract Law, see S Troiano, ‘The CISG’s Impact on EU Legislation’
[2008] Internationales Handelsrecht 221, 234–49.
227
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 288.
228
See text accompanying n 223.
229
Similarly, Troiano, ‘CISG’s Impact’ [2008] Internationales Handelsrecht 221, 241–2 in fn 160, anticipated
before the full edition of the DCFR was published that ‘the comments will shed some light on the rationale
behind this rule, which seems to be connected to the general principle of freedom of contract and to the
requirements of good faith and fair dealing in contractual transactions. It would be unfair to bind the parties to
the form which they originally agreed upon when evidence is given that at a later stage they changed their
minds reaching a new agreement, although they did not use writing.’ This at least adds the ‘freedom of
contract’ argument, which is missing in the DCFR-comments (and which was rejected above 4.). Presumably
this ‘reasoning’ – both the ‘freedom of contract’ and the ‘good faith’/unfairness arguments – would not meet
with Judge Posner’s approval, see text accompanying n 100.
230
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 289, Note no 1.
62
 The bold assertion is not even borne out if one takes the drafters’ word for what the
position is in the various countries: They list nine EU countries for the ‘evidential value
only’ position, three in which the position is ‘mixed or unsettled’, and seven EU countries –
not counting the CISG and UNIDROIT Principles and the US American UCC – in which
NOM-clauses are enforced.231
The assertion becomes even less accurate if one takes a closer look at the
jurisdictions that are cited for the ‘evidential value only’ position.
First, Germany is said to fall into this category. The discussion above should have
made clear that this is hardly an adequate description of the legal position in Germany:
while it is true that ‘simple’ NOM-clauses remain ineffective,232 individually negotiated
double NOM-clauses are regularly used and enforced in Germany in the commercial context,
and increasingly beyond.233 The complete silence on the distinction between ‘simple’ and
‘double’ NOM-clauses in the DCFR notes on Germany is slightly surprising. Every contract
law textbook and commentary on the German Civil Code deals with the distinction; and the
final sentence in the comments on DCFR II.-4:105234 appears to be directed specifically
against double NOM-clauses. A partial explanation can be found in the sources consulted on
the German position: the only reference on German law is to a commentary on the CISG,
which mentions as an aside that the German position differs from CISG Article 29(2).235 I
would submit that a commentary on the CISG is not the most authoritative source on
German law.236

231
The notes do not mention Scots law. I have not done anything worth the name ‘research’ into the
enforceability of NOM-clauses under Scots law, but a relatively recent decision by a Lothian and Borders
Sheriff Court (R&SM Nichol Ltd v PA Laing & Ing Lease (UK) Ltd 2009 WL 2221986) seems to indicate in
passing that the parties to a contract could enforce the NOM-clause: ‘I do not consider that the averments in
support of a novation agreement are necessarily irrelevant. The contention that no such arrangement could
come into existence because the original agreement did not permit variations except in writing is an argument
which I consider to be open only to the original parties to the agreement.’ (emphasis added). The Court cites
to McBryde, Contract in Scotland § 25.06. While this paragraph may be read as supporting the enforceability
of NOM-clauses, the passage is not unambiguous: ‘For Scots law the arguments will be: (1) the non-variation
clause should apply unless there is a clear indication to the contrary; (2) the parties can novate on their original
contract; their original consent is not immutable’. The ambiguity of this passage results from the undefined
extent of the proviso ‘unless there is a clear indication to the contrary’.
232
Above 2.
233
Above 3.
234
Above text accompanying n 223.
235
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 289, Note no 3, citing to P Schlechtriem in
P Schlechtriem & I Schwenzer (eds) CISG (4th German edn, Munich: CH Beck, 2004) Article 29 para 19. The
drafters of the DCFR add a citation to a 1976-edition of a commentary on the Uniform Law for International
Sales, the predecessor of the CISG (ibid).
236
Of course, some commentaries on international instruments provide sufficiently detailed information on
German law to mention the double NOM-clause case law, see Kleinheisterkamp in Vogenauer &
Kleinheisterkamp, PICC, Art 2.1.18 in fn 450. French law has met with a similar fate: The French legal
63
 Secondly, Italy is counted into the category of countries giving only evidential value
to NOM-clauses. I do not claim to have any knowledge of the Italian law on NOM-clauses.
However, the comparative notes in the DCFR themselves cite only to one court decision for
Italy – and if one looks up this decision, it did enforce the NOM-clause in question.237 This
is arguably not the best support for the claim that in Italy NOM-clauses are unenforceable.
Additionally, the comparative notes on Italy in the DCFR cite to a book comparing the
Principles of European Contract Law (PECL) with Italian law.238 If one consults the book in
question, one finds the following analysis of Italian law and its comparison to the PECL-
equivalent of DCFR II.-4:105, PECL Article 2:106:
Pursuant to Art. 1352 C.C., if the parties have agreed in writing to adopt a specified
form for the future contract it is presumed that such form was intended for the
validity of the contract (ad substantiam) and not merely for evidentiary purposes (ad
probationem). The same rule applies if parties to a contract have agreed in writing to
a specified form for any subsequent addition to and/or modification of the agreement.
In this case modifications and additions made in a different form are deemed null
and void. It follows that any modification to a written modification clause would be
made in writing (Cass. civ., Sez. III, 14 April 2000, no. 4861; but see also contra
Cass. civ., Sez. I, 5 October 2000, no. 13277). . . . As mentioned, written
modification clauses only carry evidentiary weight under the PECL. On the contrary,
Italian law recognises the enforceability of such clauses, within the limits stated
above.239
Again, this hardly lends support to the DCFR-drafters’ conclusion that NOM-clauses ‘have
evidential value only’ in Italy. Given that there appears to have been no comparative
research on Italian law beyond this one book, it is unclear to me why the drafters of the
DCFR consider Italy to support the rule suggested in DCFR II.-4:105(1). The opposite
would appear to be closer to the truth, although one could also argue that the apparently not
entirely consistent case law places Italy in the category ‘mixed or unsettled’.

position in the notes to DCFR II.-4:105 is deduced exclusively from the note written by a German
commentator in a 1976-commentary to the Uniform Law for International Sales.
237
Cass civ, Sez III, 14 April 2000, no 4861.
238
A Monti in: L Antoniolli & A Veneziano (eds), Principles of European Contract Law and Italian Law (The
Hague: Kluwer Law International 2005) 106.
239
Ibid. The ‘limits stated above’ refer first to the fact that the extent of the NOM-clause depends on the party
agreement: ‘if, for instance, the parties only agree that additions and modifications shall be made in writing,
termination by conduct . . . is admissible (See Cass. civ., Sez. III, 24 June 1997, no. 5639).’ Secondly, the
author points out that special rules apply for NOM-clauses in standard terms and conditions. Thirdly, the
author may refer to the apparently not entirely consistent case law (see the passage quoted in the text).
64
 Thirdly, the drafters of the DCFR put Greece into the category ‘evidential value
only’ with the following justification: ‘The Supreme Court of Greece has held that even if
the parties have agreed to conclude their contract in writing, they may later orally agree to
modify it . . . Even contracts for which the law requires form for their modification may be
ended by oral agreement . . .’240 All this may well be true, but it does not give any indication
as to the position on NOM-clauses – even in jurisdictions enforcing NOM-clauses, written
contracts can usually be modified orally and statutory form requirements do not prevent oral
rescission.241
Fourthly, England is said to fall into this category as well. The comparative notes in
the DCFR claim that there is no case law on the question. This was true when the PECL
were written; now, it is no longer true. The unresolved conflict between Asif on the one
hand, and World Online on the other, places England into the category ‘mixed or
unsettled’.242
On the other hand, the DCFR cites Austria as being one of the jurisdictions that do
give effect to NOM-clauses. In reality, the Austrian case law is significantly more hostile to
NOM-clauses than the German case law.243

240
C von Bar, E Clive, H Schulte-Nölke et al, DCFR Full Edition, Vol I, 289 (internal citations omitted).
241
See the references in nn 21, 151 above. I do not claim that NOM-clauses are enforced in Greece – in this
regard, I am uncertain. Parties can certainly agree on form requirements, Art 159 of the Greek Civil Code, cf
SC Symeonides, ‘The General Principles of Civil Law’ in KD Kerameus & PJ Kozyris (eds), Introduction to
Greek Law (3rd edn, Alphen aan den Rijn: Kluwer Law International, 2008) 91; I Karakostas, Einführung in
das griechische Privatrecht (Athens: Ant N Sakkoulas, 2003) para 91. Pursuant to Art 159(2) of the Greek
Civil Code, juridical acts not conforming to an agreed form requirement are null and void (only) in cases of
doubt. Whether there is a ‘doubt’, however, depends on the parties’ will (Symeonides, ibid) – in other words,
where the parties agree that modifications have to be in writing to be valid, as in a typical NOM-clause, this
appears to be enforceable. The legal provision thus appears to establish the same rule as German law (see n 38
above). It is not entirely clear, however, whether Greek courts interpret Art 159(2) of the Greek Civil Code in
the way German courts treat ‘simple’ NOM-clauses (above 2, ie arguing that a subsequent oral agreement will
supersede the NOM-clause), or in the way German courts treat ‘double’ NOM-clauses (above 3, ie enforce the
NOM-clause). Symeonides (ibid) notes that even if the parties wanted informal juridical acts to be null and
void, ‘the nullity may be cured by subsequent knowing performance’ (emphasis added). This might indicate
that only subsequent knowing performance will cure the nullity, so that a mere subsequent oral agreement to
modify would not supersede the NOM-clause. It could, however, also be the case that the subsequent knowing
performance is only mentioned as an example, pars pro toto. Not being able to do research in Greek, I cannot
decide which of these interpretations is the correct one. All I am claiming is that the justification advanced in
the notes to DCFR II.-4:105 with respect to Greek Law is a non sequitur.
242
See above 3 at nn 72–9.
243
While § 884 of the Austrian Civil Code (ABGB) sounds as if an agreed form requirement were to be given
effect, Austrian case law appears to argue along the same lines as German courts with regard to ‘simple’
NOM-clauses – the NOM-clause is held to be abrogated by any modifying agreement between the parties. In
Austria this is even extended to ‘double NOM’ clauses. I have not gone to great lengths to verify the current
Austrian position; but sources I have at hand indicate that any modifying agreement will supersede the NOM-
clause. See Österreichischer Oberster Gerichtshof (OGH) 16 April 2009, 2 Ob 221/08a, and OGH 7 March
2006, 5 Ob 37/06m (available via <http://www.ris.bka.gv.at/Judikatur/>), both holding that even where the
parties have agreed on what would under German law be called a double NOM-clause, they are free to modify
the contract informally; obiter referred to with approval by OGH 2 September 2009, case 7 Ob 136/09i
65
 From the DCFR-list of EU countries in which NOM-clauses are not enforced, this
leaves the Scandinavian countries, Slovenia, and Ireland. 244 While these are no doubt
important jurisdictions, it is not entirely convincing to say that they constitute ‘most of the
countries of the Union’ of 27 Member States (and at least 28 jurisdictions), even if one
should have to add Austria and the Netherlands to this list. Furthermore, it may be worth
noting that the Scandinavian countries and Slovenia – as well as the Netherlands and
Austria – are CISG Contracting States and therefore enforce NOM-clauses where the CISG
is applicable.245
More importantly, where courts, and not legislators, have rejected the enforceability
of NOM-clauses, the decision may not have been based on a deliberate policy-decision
against NOM-clauses. Instead, the courts may have felt that in the absence of a clarifying
statutory provision they have to enforce the informal modifying agreement for reasons of
‘freedom of contract’ (at the time of the conclusion of the modifying agreement, t2). To the
best of my knowledge, no court has ever objected to a statutory provision making NOM-
clauses enforceable. The World Online-Court in England, for example, expressed
considerable sympathy for the position that NOM-clauses should be enforceable, but
ultimately felt that, in the absence of a provision such as UCC § 2-209, the principle of the
freedom of contract in t2 should prevail. Court decisions about the legal position in the
absence of a statutory provision are not necessarily good guidance for the question whether
a legislator should make NOM-clauses enforceable. A legislator is free to decide on the
solution supported by the best policy arguments – within the bounds of constitutional law –,
whereas a court is constrained by statutory and case law.
The question whether ‘most’, ‘many’, or ‘some’ countries in the Union enforce (or
do not enforce) NOM-clauses is therefore largely irrelevant. Comparative law is not about
counting jurisdictions. A comparative analysis should not turn on the number of
jurisdictions supporting a particular rule, but on the merits and demerits of the respective

(available via <http://www.ris.bka.gv.at/Judikatur/>); see also OGH 9 February 1995, 8 Ob 16/94, 68

Entscheidungen des österreichischen Obersten Gerichtshofs in Zivilsachen 153, 163 (no 28);
Oberlandesgericht Innsbruck 18 December 2007, case 1 R 273/07t,
<http://cisgw3.law.pace.edu/cases/071218a3.html> (Steel bars case): ‘The parties are free to choose a different
form at any time according to the principles of private autonomy. . . [I]t is possible to deviate from an agreed
written form at any time by mutual explicit or tacit consent. . . This does also apply to the present case where
the parties have agreed that written form is required in order to deviate from the agreement on written form (5
Ob 37/06m with further references)’ (translation by D Nagel); P Rummel in P Rummel (ed), Kommentar zum
Allgemeinen bürgerlichen Gesetzbuch Vol 1 (2nd edn, Manz: Vienna 1990) § 884 paras 2 and 3.
244
For the avoidance of doubt: I have not done any research into the Scandinavian, Irish, or Slovenian position.
245
The Scandinavian states have entered CISG Art 92-reservations only to Part II of the CISG; and Art 29(2)
is in Part III of the CISG (hence the Scandinavian attempts to move Art 29(2) into Part II of the CISG, see
above n 217). Intra-Scandinavian trades are, however, excepted due to an Art 94-reservation.
66
positions. There are policy reasons both for and against the enforcement of NOM-clauses;
but the DCFR regrettably does not embark on this discussion at all. Nor is it a defence that
the DCFR is not the work of a legislator, but a purely ‘academic’ exercise. If the DCFR is
an academic work, it should develop the policy arguments for both sides, research the legal
position in the various Member States, and leave the decision on the rule to be adopted to
others;246 this is not what the DCFR has done.

DCFR II.-4:105(2): An Exception to a Rule that Does Not Exist

A further complication is added by DCFR II.-4:105(2), which establishes a reliance-based
exception. The reason for this exception is difficult to grasp. As discussed above, the
drafters of the DCFR have opted for the position that NOM-clauses have no effect; instead,
any informal agreement to modify the contract or terminate the contractual relationship, if
sufficiently proven, will do. To this, DCFR II.-4:105(2) adds a reliance-based exception.
Such reliance-based exceptions exist, and are important, in jurisdictions that do enforce
NOM-clauses.247 And indeed, DCFR II.-4:105(2) is modelled on these exceptions.248 There
is no explanation why such an exception should be necessary if the NOM-clause is
unenforceable in the first place.
In order to resolve the contradiction, one could hypothesize that the DCFR in II.-
4:105(1) makes a fine distinction between an informal agreement to disapply the NOM-
clause and an informal agreement to modify the contract (or terminate the contractual
relationship): If this were what DCFR II.-4:105(1) meant, then while an informal agreement
to disapply the NOM-clause would suffice, such an agreement would also be necessary
before the contract could be modified informally. Such a distinction can occasionally be
found in German case law and literature on ‘simple’ NOM-clauses,249 and the following
passage from the comments to DCFR II.-4:105(2) may lend support that this is what the
DCFR-drafters meant:
If the parties have reached an oral agreement – for example, they have agreed orally
to modify a contract that contains a merger clause or a ‘no oral modification’ clause
– but it cannot be shown that they have agreed to disapply the clause, yet one party

246
Also cf House of Lords European Union Committee, European Contract Law, 25 [80] (‘One general point
that can be made about the detailed drafting is that the DCFR does not offer a European legislator choices’).
247
Above 8, discussing UCC § 2-209(4),(5), California Civil Code § 1698(d), CISG Art 29(2) (second
sentence), UNIDROIT Principles Art 2.1.18 (second sentence), and the ‘good-faith’ defence in Germany.
248
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 291 Notes nos 18, 19.
249
Above nn 49, 50.
67
 has reasonably acted in reliance on the oral agreement, the other party will be
precluded from invoking the clause.250
While this would explain the presence of the reliance-based exception in DCFR II.-4:105(2)
– for cases in which only the agreement to modify, but not the separate agreement to
disapply the NOM-clause can be proven –, this interpretation is incompatible with the
comments on DCFR II.-4:105(1), which, it should be recalled, state: ‘If, therefore, it can be
shown that both parties agreed to a modification of the contract terms or a termination of
the contractual relationship, but did not use the specified form, effect must be given to their
agreement.’ This clearly indicates that the modification of the contract terms (or a
termination) is sufficient in itself, without a requirement for any separate agreement to
disapply the NOM-clause. Also, if the DCFR wanted to distinguish the informal agreement
to disapply the NOM-clause from the informal agreement to modify (or terminate), it would
have to clarify what the requirements for such a separate agreement should be: if not every
informal agreement to modify or terminate is to be seen as an implicit agreement to disapply
the NOM-clause, one would have to demand (i) that the agreement to disapply the NOM-
clause must be explicit, or at a minimum (ii) that the parties must have thought of the NOM-
clause when agreeing on the modification. The DCFR comments to paragraph (1) appear to
require neither.
What is more, the illustration given for the scope of application of DCFR II.-4:105(2)
does not support the hypothesis that the DCFR makes the fine distinction between informal
agreements to disapply the NOM-clause and the informal agreement to modify or terminate.
For fear of misstating the DCFR-drafters’ position, I will quote the illustration and its
proposed solution verbatim:

Illustration
A construction contract contains a clause providing that ‘this contract may only be
modified in writing signed by both parties’. Subsequently the parties orally agree to
some changes in favour of the owner. The changed obligations are performed. When
later the contractor invokes another oral modification made in its favour the owner
invokes the ‘no oral modification clause’.
The contractor may invoke the performance of the obligations as modified by the first
oral agreement to show that the second oral agreement, in favour of the contractor, is

250
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 288, comment B.
68
 binding on the owner. The contractor has in fact relied on the abrogation of the ‘no
oral modification’ clause.251

I must admit that I am puzzled by this application of the reliance-based exception. As the
comments on DCFR II.-4:105(1) and the first set of oral modifications in the illustration
show, any oral modification of the contract will be valid despite the NOM-clause. 252
Accordingly, the contractor could simply prove the existence of the ‘second oral agreement’
under the general rule of DCFR II.-4:105(1) – I fail to see why the contractor should feel the
need to resort to any reliance-based ‘exception’ at all.

Is the Criticism of DCFR II.-4:105 Representative for the DCFR?

It should be clear by now that I am not particularly impressed by the depth of the
comparative analysis of NOM-clauses in the DCFR. Perhaps I am unduly harsh in my
criticism of DCFR II.-4:105. I have picked on one particular provision, perhaps not even a
very important one, and looked at less than a handful of jurisdictions in some detail. In
contrast, the drafters of the DCFR had to worry about more than 27 contract law regimes
and more than 250 provisions in the first three books of the DCFR alone. The DCFR-
drafters had a Herculean task to manage, and they have done an impressive job in a very
short period of time.
Also, I do not mean to suggest that similar deficiencies affect all, most or even many
other DCFR provisions. It may even be that DCFR II.-4:105 is the lone outlier.
And yet, the provision that I have singled out for criticism is not picked completely
at random. It is a DCFR-provision that consciously deviates from its counterparts in the
CISG and the UNIDROIT Principles. This should have put the drafters on guard. While
there is certainly no guarantee that the solutions in the CISG and the UNIDROIT Principles
are necessarily the best ones, the drafters of these instruments usually had their reasons for
adopting particular solutions. One should at least have a close look at these reasons before
rejecting them. This was not done by the drafters of DCFR II.-4:105.
The deficiencies detected in the comparative analysis on NOM-clauses might have
been prevented if the multi-jurisdictional team working on the DCFR had required an expert
from each jurisdiction to look at the rules from her or his jurisdiction. No single comparatist

251
von Bar, Clive, Schulte-Nölke et al, DCFR Full Edition, Vol I, 288–9.
252
Except if the fact that the first set of modifications was performed were to be a requirement for their
becoming binding; but nothing in the model rule or in the comments indicates that this should be the case.
69
could conceivably do reliable legal research in the more than 27 contract law regimes in the
European Union. I do not blame any non-German lawyer for not identifying the complicated
case law on double NOM-clauses; nor any non-English lawyer for overlooking the few,
unreported precedents in England that were handed down after the publication of the PECL;
nor any single researcher for any of the other discrepancies identified here.253 The problem
lies in the division of labour along provisions, instead of along jurisdictions. Each
comparatist can usually be expected to work confidently in between two and, say, ten
jurisdictions. Beyond this, legal research becomes superficial, often because of language
barriers, but – even if language is not an obstacle – because of a lack of sufficiently broad
and deep knowledge of the legal systems and the respective research methods used in the
jurisdictions. For individual comparative research this means that either the researcher has
to confine herself to the jurisdictions in which s/he is competent, or that the researcher has
to resort to slightly superficial research in some of the jurisdictions.254 For comparatists
working as a team, tertium datur.255 The teams of outstanding comparatists working on the
DCFR were large and well funded, 256 and they included luminaries from across the
European Union. It should have been possible to have one senior researcher from each of
the various jurisdictions work in parallel on each of the DCFR provisions.
Perhaps one of the problems was also the decision to avoid individual identification
of the drafters of each provision and its notes and comments; the Acquis Group’s work
shows that such identification is possible.257 It is one thing to acknowledge the contribution
by the entire team for the whole project. Yet not identifying the individual authors for the
separate provisions diffuses responsibility and may not provide the best incentive to elicit
the maximum individual effort. This is particularly true if the team of researchers includes
young researchers, who will presumably not have a great interest in investing the best years
of their life into work that is not attributed to them personally and which will thus have
limited reputational value; and even more so, if the turnover of researchers is high. None of
this is to deny that the vast majority of the researchers have been extremely dedicated to

253
After all: ‘[T]he cleverest comparatists sometimes fall into error; when this happens, the good custom of
workers in the field is not to hound the forgivable miscreant with contumely from the profession, but kindly to
put him right.’ Zweigert & Kötz, Introduction to Comparative Law 36.
254
My own attempt here is an imperfect combination of both these approaches.
255
cf Zweigert & Kötz, Introduction to Comparative Law 42.
256
cf von Bar, Clive, Schulte-Nölke et al, DCFR Outline Edition, 47–54 and 55–6 respectively. See also von
Bar, ‘Coverage and Structure’ [2007] European Review of Contract Law 350, 357: ‘[A]s we had the research
facilities to go into these questions in detail we decided to cover non-contractual obligations as well.’ This
suggests that the resources for a thorough revision of the PECL would have been available.
257
Acquis Group, Contract II, xxi–xxii.
70
their work; but as an institutional set-up, it is suboptimal. Economists immediately spot a
free-riding problem.
Another problem of DCFR II.-4:105, which may affect other DCFR provisions as
well, is that it does not sufficiently distinguish between commercial B2B, B2C, and C2C
relationships.258

11. Conclusion
The comparative work done on NOM-clauses in the DCFR is unsatisfactory.259 A drafter of
any binding or non-binding instrument should start ‘from scratch’ and ask the following
questions:

(i) Should NOM-clauses be given effect?

This is a policy question, and of course a drafter can decide it either way.260 If the answer is
no, then the drafter should say so clearly, and not couch it in diplomatic language of a
presumption that is always rebutted. The better policy reasons, however, support an
affirmative answer,261 although a general rule of enforceability will require to be qualified
by exceptions.262
(ii) Should NOM-clauses only be effective when individually negotiated or should a
NOM-clause in standard terms and conditions be enforced?
On the one hand, most NOM-clauses in practice are contained in standard terms and
conditions, and giving effect only to individually negotiated clauses would render nugatory
all these NOM-clauses. On the other hand, the only policy concern against NOM-clauses is
arguably that they may catch one of the parties unawares. Therefore, there is an argument to
be made that only individually negotiated clauses should be enforced; alternatively, one
could opt for an intermediate solution, such as the one chosen by UCC § 2-209(2), ie
requiring a separate signature acknowledging the NOM-clause. Given that in practically all
of the cases involving NOM-clauses, it was businesses that were apparently caught unaware
of the NOM-clause, one may consider going further than UCC § 2-209(2) and require a

258
This is a frequent criticism of the DCFR, cf, eg, House of Lords European Union Committee, ‘European
Contract Law’, 24 [80].
259
Above 10.
260
Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449 (1988), passim; idem, ‘Commercial
calamity’, (2007) Ohio State Law Journal 335, 347–8.
261
Above 4.
262
Above 5–9.
71
separate signature even for transactions ‘between merchants’ (B2B relationships).263 The
question whether to enforce NOM-clauses in standard terms and conditions should be seen
in context with the next question:
(iii) Should NOM-clauses only be available to businesses?
The answer depends on the perceived dangerousness of NOM-clauses, and on one’s positive
and normative conception of the rationality of consumers. The policy arguments for
enforcing NOM-clauses are undoubtedly the strongest for B2B relationships,264 while for
B2C relationships they are undoubtedly the weakest.265
(iv) Should novations be subjected to an effective NOM-clause?
Because the distinction between ‘modifications’ and ‘novations’ is a tenuous one, the rule
should subject rescission, termination, discharge and novation to the NOM-clause as well.266
The level of detail into which the drafter will have to go here depends on the culture of
statutory interpretation: In a civil law codification, a rule such as CISG Article 29(2) would
probably suffice, whereas the strict interpretation traditionally applied to common law
statutes may require a much more detailed provision such as New York General Obligations
Law § 15-301.267
(v) Should collateral agreements be subjected to an effective NOM-clause?
While the possibility of collateral agreements may weaken the effectiveness of NOM-
clauses, this is hard to prevent. The drafter could, however, clarify that informal collateral
contracts cannot contradict or qualify a contract containing a NOM-clause.268
(vi) Should there be exceptions?
If one chooses to enforce NOM-clauses (and only then), one will have to give thought to
possible, eg reliance-based, exceptions. Theoretically, the ‘cleanest’ solution would be to
disregard any waiver/estoppel/good faith arguments, and require the modification,
termination or novation to be in writing in all cases.269 Only this rule – we could call it the
‘Draconian NOM’ provision – ensures with a very high probability that the NOM-rule can

263
Similar Bergsten, ‘Amending the Contract’, 60: ‘[T]here is some concern that such clauses in general
conditions may not be thought about or be appropriate when the individual contract is entered into. While this
concern is most prevalent in regard to contracts between a merchant and a consumer, the concern exists in
regard to commercial contracts as well.’
264
cf the quote at n 124.
265
cf the quotes at n 136 and in n 263.
266
Above 6.
267
See above n 170. I doubt that US American courts would today apply the same narrow interpretation to a
statute that they applied in Green v Doniger.
268
Above n 173.
269
cf Hillman, ‘Article 29(2)’, 21 Cornell International Law Journal 449, 463 (1988), suggesting that
‘[p]erhaps future drafters should abandon compromise, which inevitably generates confusion. Instead, they
should swallow hard and then decide simply to enforce NOM clauses or to abolish them. . .’
72
unfold all its beneficial effects by preventing oral evidence about informal attempts at
modification. However, experience in many legal systems demonstrates that even statutory
form provisions require a safety valve to prevent inequitable results in extreme cases. The
Draconian NOM-clause, while making things conceptually cleaner, would not be well suited
to deal with the vagaries of life.270 It may be prudent for any statutory exception to stress its
exceptional status, and perhaps to provide more detail on what can be considered
‘reasonable reliance’, so as to prevent an interpretation that renders the rule nugatory.271

The DCFR-provision, unfortunately, does not discuss these policy issues and the various
options.272 Instead, it tries to ossify a legal position that has ‘virtually nothing to commend’
itself,273 without stating any substantive reason and without backing up its strong claims
about the comparative justification of the rule by thorough research. If DCFR II.-4:105 is
any indication for the depth of comparative research on the other provisions in the DCFR,
then a thorough re-evaluation of the DCFR is called for. The DCFR has opted for breadth
instead of depth. Given the time constraints, the choice between breadth and depth may
have been inevitable.274 The best choice would be to opt for limitations neither in depth nor
in breadth, but simply to remove the artificial time constraints on the project.275

270
For a similar assessment, see Snyder, ‘Concept of change’ [1999] Wisconsin Law Review 607, 655–70; see
also JS Ziegel in idem & C Samson, Report to the Uniform Law Conference of Canada on Convention on
Contracts for the International Sale of Goods (July 1981), available at
<http://www.cisg.law.pace.edu/cisg/wais/db/articles/english2.html>: ‘The second sentence of art. 29(2)
introduces an important, and very necessary, qualification to the rule in the first sentence. . . No doubt there
may be arguments about the “extent” to which the other party has relied “on that conduct” in a particular case,
but this difficulty seems to me unavoidable.’
271
See Hillman, ‘Commercial Calamity’, 68 Ohio State Law Journal 335, 348.
272
See above n 246.
273
See above n 134.
274
The frustration of being caught between the need to adhere to the strict deadlines and the calls for deeper
research is expressed by one of the most prominent members involved in the drafting of the DCFR: ‘It is a
unique experience, however, to work under the utmost time pressure generated by one Commission
department while at the same time other departments which were not involved in the negotiations here and
there leave one with the impression that the worst we can do is to fulfill our contract!’ (von Bar, ‘Coverage and
Structure’ [2007] European Review of Contract Law 350, 356); also cf C von Bar & H Schulte-Nölke,
‘Gemeinsamer Referenzrahmen für europäisches Schuld- und Sachenrecht’ [2005] Zeitschrift für Rechtspolitik
165, 167.
275
See Zimmermann, ‘Present State’ 57 American Journal of Comparative Law 479, 493–4 (2009): ‘What is
to be very much regretted is that what began as a genuinely academic exercise . . . has been transformed into a
political project with all the concomitant features that this entails. The most damaging among them . . . has
been the extreme pressure of time. . . [A] cooling-off period would be desirable, and the European contract law
project should be taken off the political track.’; Mance, ‘Civilising the Common Law’ [2007] EBLR 77, 97:
‘[T]he time is not yet ripe for wholesale codification . . . [T]here is a need for a structured, evolutionary
approach.’; Eidenmüller et al, ‘Common Frame of Reference’, (2008) 28 Oxford Journal of Legal Studies 659,
708: ‘The debates in the years to come, and the competition for the best arguments and solutions, will reveal to
what extent any of these texts [scil.: in the DCFR] can gain authority imperio rationis . . . These debates
require considerable time. The DCFR clearly demonstrates that the time has not yet come for a politically
legitimated text.’
73