Grammaticalization in Morphology: 1.1 Grammaticalization As A Phenomenon
Grammaticalization in Morphology: 1.1 Grammaticalization As A Phenomenon
Grammaticalization in morphology
Muriel Norde
gradually turn into grammatical items (such as auxiliaries or pronouns), after which they may
continue to evolve into yet more abstract function words or even inflectional affixes. It is a
reductive process, characterized by loss of semantic and phonological substance, as well as loss
/ or discourse, either simultaneously or in succession (Norde 2012, Norde & Beijering 2014;
Well-known examples of grammaticalization include the change from motion verbs into
future auxiliaries (English to be going to, French aller, Dutch gaan), the development of an
indefinite article out of the numeral ‘one’ (English a(n), French un(e), Dutch een), or the shift
from demonstrative pronoun to complementizer (English that, French que (< Latin quod),
Dutch dat). Such changes are assumed to proceed along the so-called ‘cline of grammaticality’
(1) content item > grammatical word > clitic > inflectional affix
It goes without saying that the transitions between the different points on the cline are very
diverse. This has led some authors (e.g. Detges & Waltereit 2002: 188) to restrict the term
‘grammaticalization’ to the shift from content item to grammatical word, whereas others (e.g.
Givón 1991: 305; Traugott 2002: 26-27) distinguish between ‘primary grammaticalization’ and
‘secondary grammaticalization’, although these terms have been defined in different ways (see
further 2.2).
regularity – the examples just mentioned do not only occur in English, French and Dutch, but
in a range of other (typologically unrelated) languages as well. This regularity can be explained
as the result of cognitive processes that are not language-specific (Heine 1997). For instance, it
is commonly observed that spatial notions are conceptualized using the human body as a point
account for shared etymological pathways, such as locative adpositions meaning ‘behind’
developing out of a noun meaning ‘back’ (e.g. Danish bag ‘behind’ < Old Danish baker ‘back’),
or temporal adpositions developing out of locative ones (e.g. English before, after, between).
Research into grammaticalization phenomena can be traced back to at least the eighteenth
century (e.g. Condillac 1749; for more references see Heine, Claudi & Hünnemeyer 1991: 5ff.)
and the term ‘grammaticalization’ (grammaticalisation) was coined by Meillet in 1912 to refer
grammatical character to an erstwhile autonomous word ].i The study of these phenomena was
in its heydays from the late 1980s until the late 2000s, which saw a steady flow of textbooks
(Heine, Claudi & Hünnemeyer 1991; Lehmann 1995 [1982]; Hopper & Traugott 2003) and
collective volumes, both conference proceedings and journal special issues (Traugott & Heine
1991a,b; Bybee, Perkins & Pagliuca 1994; Giacalone Ramat & Hopper 1998; Fischer,
Rosenbach & Stein 2000; Bisang, Himmelmann & Wiemer 2004; Norde, Lenz & Beijering
2013; Von Mengden & Simon 2014; Breban & Kranich 2015). Other important reference works
are the OUP handbook (Narrog & Heine 2011), and a lexicon (Heine & Kuteva 2002). In
addition, grammaticalization has been the theme of a series of dedicated conferences, New
Reflections on Grammaticalization, which ran from 1999 to 2012, and lead to a further series
of collective volumes (Wischer & Diewald 2002; Fischer, Norde & Perridon 2004; López
Couso & Seoane 2008 / Seoane & López-Couso 2008; Davidse, Breban, Brems & Mortelmans
Although grammaticalization changes naturally did not vanish from the historical
linguistics agenda, the first decade of this century did witness a paradigm shift. Recent
construction grammar (e.g. Goldberg 1995, 2006; Croft 2001), in particular diachronic
construction grammar (Noël 2007; Bergs & Diewald 2008; Barðdal, Smirnova, Sommerer &
Gildea 2015). Because constructions are defined as symbolic pairings of form and meaning, a
which involve changes in both form and meaning, but the objectives of these two approaches
do not completely overlap. Whereas both are concerned with the lexical origins of grammar
and grammatical change, grammaticalization research places strong focus on the cognitive
construction grammar highlights aspects of change that had remained somewhat under-exposed
in grammaticalization research (see Cuyckens 2018 and Coussé, Andersson & Olofsson 2018
fully schematic construction such as the ditransitive construction), which are organized in a
2007, 2010, 2013), which allows for quantitative methods, primarily corpus linguistics (Hilpert
2013, 2015). After pioneering work by Traugott (2007, 2008, 2014) and Trousdale (2008,
2012), the publication of two joint works (Traugott & Trousdale 2010, 2013) laid the
constructional networks.
1.3 Criticism
linguistics for about three decades, it was also the subject of much criticism, starting with
Newmeyer 1998 and, in particular, the papers in Campbell 2001. The criticism was primarily
aimed at two aspects, first, that grammaticalization is a separate process, and secondly, that
grammatical change is unidirectional. As to the former point, Joseph (2001) has argued that all
attested independently and can be accounted for in their own terms, so there is no need for a
new concept. Similarly, Newmeyer (1998: 235) asserts that grammaticalization is merely an
epiphenomenal result of changes on different levels (such as phonology and semantics), that
just happen to occur simultaneously or in short succession. This line of reasoning was however
rejected by grammaticalization researchers, e.g. Heine & Kuteva (2002: 2f.), who point out that
these changes do not simply happen to co-occur, but are interrelated. Much earlier, Bybee,
Perkins & Pagliuca (1994: 298), had convincingly argued that ‘The processes that lead to
grammaticalization occur in language use for their own sakes; it just happens that their
grammaticalization, but to language change generally. In the early days of theorizing about
grammaticalization (e.g., Heine, Claudi & Hünnemeyer 1991, Traugott & Heine 1991c, for
further discussion and references see Norde 2009: 48ff.), there was a firm belief that
grammaticality, given in (1) above. This was not just a matter of definition, it was also held that
change in the opposite direction was simply precluded. As Lehmann (1995 [1982]: 19)
emphatically put it: “no cogent examples of degrammaticalization have been found”. The
interesting effect of this quote is that Lehmann had unintentionally created a Popperian black
swan and turned unidirectionality into a falsifiable hypothesis. As a result, others started to
search for, and found, these black swans, i.e. cases of degrammaticalization (e.g. Ramat 1992;
Norde 1997; Janda 2001; Van der Auwera 2002; Willis 2007; see Norde 2009 for a survey of
the earlier case studies). Although not all of these examples were eventually accepted as
2 Classifying (de)grammaticalization
2.1 Introduction
but the effects depend on the kind of (de)grammaticalization involved. This section provides a
brief account of these different kinds of changes – specific examples are given in the next
number of respects, they also have a few properties in common, which delineate them from
other diachronic processes (Norde 2011: 476f.). First of all, both grammaticalization and
degrammaticalization are ‘composite changes’, i.e. the concurrence of ‘primitive’ changes on
different levels (semantics, morphology, syntax, and/or phonology), that can be described in
terms of Lehmann’s (1995[1982]) parameters (see Norde 2012). Secondly, both are gradual, in
the sense that it is not necessarily the case that all primitive changes occur simultaneously.
in novel grams or structures. This restricts the number of potential degrammaticalization cases
– when a lexical item acquires a grammatical function and then loses it again, leaving only the
original lexical meaning, the change will not qualify as a case of degrammaticalization, but as
an instance of ‘retraction’ (see further Haspelmath 2004: 33ff.). Lastly, both are context-internal
changes, in the sense that “the identity of the construction and the element’s place within it are
always preserved” (Haspelmath 1999: 1064). Initial stages of grammaticalization typically lead
to what Heine (2002: 91ff) has termed ‘bridging contexts’, i.e. ambiguous contexts in which
the grammaticalizing element can be interpreted in two ways. For example, in the bridging
context the day following his intervention, following can be interpreted either as a gerund (‘the
day which followed his intervention’) or as a preposition (‘the day after his intervention’), its
referring to the early stages (from lexical item to grammatical item) and subsequent stages (from
grammatical item to another grammatical item, or from grammatical item to bound morpheme)
respectively. Basically, the terms refer to the two types of change identified in Kuryłowicz’s
(1975 [1965]: 52) often-cited definition: “Grammaticalization consists in the increase of the
grammaticalization and the relation between the two. The present paper follows the definitions
morphosyntactic contexts of constructions and lexical categories into functional categories” (p.
lexical verb with volitional meaning, it developed into a future auxiliary will (primary
grammaticalization).
(2) lexical verb > auxiliary > enclitic auxiliary > inflectional tense marker
└─────┬─────┘└───────────────┬───────────────┘
Like all terms ending in ization, secondary grammaticalization has been at the heart of
definitional debates (Norde & Beijering 2014). Von Mengden (2016: 134) even suggests that
definition of grammaticalization, that is, the “product of the various turns and notional
Smirnova (2015), who argues that secondary grammaticalization is not always the continuation
of primary grammaticalization (as (2) might suggest). Not only did she find a great deal of
regularity across different secondary grammaticalization changes, these changes are not
necessarily steered by the semantics of the lexical element at the very beginning of the chain.
morphological properties may be more prominent than in less inflected ones (including
Smirnova’s and Breban’s objections are duly noted, but since this paper focuses
between the two than the direction of change. They also differ with respect to frequency and
hand, is far less commonly attested and changes are often restricted to a single language or a
grammaticalization in that it does not form chains, that is, a degrammaticalization change (e.g.
from affix to clitic) is usually not followed by a subsequent degrammaticalization change (e.g.
from clitic to independent function word), although North Saami (-)naga ‘stained with; stain’,
is establishing the direction of the change on the cline of grammaticality in (1). If the change
comprises a change from right to left on this cline, and if it is furthermore a gradual, context-
morphosyntactic properties which are typical of that word class, and gaining in semantic
‘deinflectionalization’ or ‘debonding’, depending on the input and output of the changes. Thus,
linguistic context gains a new function, while shifting to a less bound morpheme type” (Norde
2009: 152). The term debonding, finally, denotes “a composite change whereby a bound
morpheme in a specific linguistic context becomes a free morpheme” (Norde 2009: 186).
This chapter discusses a few illustrative examples of morphological changes in each of the five
of grammaticalization in the world’s languages, a full survey is naturally out of the question
(many more examples are to be found in the reference works mentioned in 1.2). Furthermore,
for reasons of space, diachronic details need to be kept to a minimum. It should be stressed,
therefore, that what are called ‘changes’ in this paper are actually ‘diachronic correspondences’,
an important distinction introduced by Andersen (2001). Andersen correctly notes that language
historians often speak of changes when it would be more appropriate to speak of diachronic
correspondences, which are the results of change. A diachronic correspondence, then, is the
synchronic states in a linguistic tradition (…)”. (Andersen 2001: 28), whereas changes are “the
historical events in a linguistic tradition by which practices of speaking vary over time”.
Changes can be observed when they are in progress, even though they often go unnoticed by
here, but with the proviso that the discussion is largely restricted to input and output of change.
major category to a minor category (Hopper & Traugott 2003: 106ff.). From a functional
perspective, decategorialization refers to the loss of discourse autonomy (Hopper 1991: 30).
For instance, when nouns grammaticalize, they lose the ability to identify participants in
discourse – top in on top of the refrigerator does not necessarily refer to the upper part of the
refrigerator. The expression can also be used when the refrigerator is lying on its side and
something else is lying on top of it (DeLancey 1994: 4). In languages such as English, where
major categories are inflected and most minor categories (with the exception of pronouns) are
inflections and / or the ability to take articles and modifiers.iv For instance, in the relational
phrase on top of, the noun top cannot pluralize nor can it take an article. Similarly, when full
verbs grammaticalize into auxiliaries they may lose person endings (cf. English I want to, he
wants to vs I wanna, he wanna (*wannas); Krug 2000: 160) and become ungrammatical in
certain contexts (e.g. certain temporal clauses such as English *Let’s wait till she will join us
Ramat (2001: 398) argues that it may be taken to mean ‘loss of categorial status’, so instead he
proposes the term ‘transcategorization’. However, as Brinton and Traugott (2005: 25) point out,
this term includes ‘lateral shifts’ (see 4.4) between major categories of the type hand (N) > to
hand (V), and is hence broader in scope than decategorialization. Although this is essentially
correct, the term ‘decategorialization’ remains less than ideal, since loss of inflection is also
found in some shifts from one minor category to another, e.g. when subordinators develop from
pronouns (Cristofaro 1998: 82f.). Conversely, loss of inflections may occur when a minor
category is subject to semantic change, as in the case of the German modal auxiliary wollen
‘want; may’, that lacks a preterite form when the verb has epistemic meaning (Diewald 1997:
24ff.).
the development of the Old Swedish neuter noun mot ‘meeting’ (Norde 2012: 77ff.). As a noun,
mot can be modified by an adjective and it is inflected (3)). In Old Swedish texts from the same
period, mot is also found in phrasal adpositions meaning ‘against, towards’, preceded by i ‘in’
or ‘on’. Initially, mot is still inflected, as in (4), but in later texts, the inflection is lost (see (5)).
At the final stage, with the preceding preposition gone as well, mot functions as a preposition
‘he shall make it known at the first encounter (i.e. to the first person he meets)’
development of quantifiers from nouns, such as English a lot of (Traugott 2008), Dutch massa’s
(from the plural of the noun massa ‘mass’; De Clerck & Colleman 2013), Italian un sacco di ‘a
bag of’ > ‘a lot of’ (Giacolone Ramat 2018), or the evolution of a 3SG indefinite pronoun from
a noun meaning ‘man’ (French on < Latin homo, German man < Mann).
In secondary grammaticalization, the consequences for morphology are not the loss of
free grammatical morphemes, often followed by full affixation. This frequently gives rise to
new inflectional categories in the nominal or verbal domain, with fully-fledged inflectional
paradigms as a result.
One such cline is the development from free demonstrative to affixal determiner in the
Scandinavian languages (Larm 1933, 1936, Faarlund 2007b, Börjars & Harries 2008, Norde
2012: 78ff., Stroh-Wollin 2015). The change is illustrated here with examples from Swedish,
where the Proto-Norse demonstrative hinn, which was postpositive, became cliticized to the
noun and later became an affix. As the cline in (7) (adapted from Stroh-Wollin 2015: 13) shows,
bound (h)inn retained its inflections, with the result that definite noun forms were inflected
twice in Old Swedish (see Table 1). This double inflection is no longer found in Modern
Table 1: Demonstrative hin and bound -in in Old Swedish (standardized forms only; see further Noreen
1904: 391, 401, 408ff.)
Secondary grammaticalization in the verbal domain includes the emergence of the Germanic
‘weak’ preterite suffixes from Proto-Germanic auxiliary *dōn ‘to do’ (e.g. Sczcepaniak 2011:
112ff.), and of new inflectional futures, e.g. in the Slavic (e.g. Andersen 2006, 2008) and the
Romance languages (e.g. Lausberg 1972, Fleischman 1982, Pinkster 1987, Roberts & Roussou
2003: 48ff.), where the Latin auxiliary habere ‘to have’ gradually fused with the verbal stem,
However, secondary grammaticalization does not always lead to in new inflectional categories
– the result may also be a single suffix. For example, the !Xun diminutive marker –mà deriving
from a relational noun mà ‘(own) child’ lost both its lexical semantics and morphosyntactic
properties (Heine 2011: 702). Likewise, the Lezgian evidential suffix -lda grammaticalized
from a single verbal form, luhuda ‘one says’. The change is illustrated in (8), which also features
the lexical verb luhu from which the evidential marker derives (Haspelmath 1993: 148).vi
zarar ja-lda.
harm COP-EVID
It also ought to be noted that suffixes do not always originate in clitics. An alternative source
are derivational affixes such as English adjectival –ly (ultimately from a Proto-Germanic noun
*lika- ‘body’), as in lovely, that developed further into an adverbial suffix, as in quickly (Brinton
bound morphemes. A well-known example is the development of English gonna from the future
auxiliary (to be) going to (itself a case of primary grammaticalization from the motion verb to
go). This is an instance of reduction and fusion whereby morpheme boundaries are lost entirely,
even to the extent that gonna is becoming increasingly ‘emancipated’ from its source going to
Degrammation involves both semantic change (from grammatical content to lexical content)
and morphosyntactic change (from minor to major word-class, typically acquiring the inflection
only rarely attested – presumably because its target is a major word class, which typically
inflects. This implies that in order for a grammatical element to degrammaticalize into a major
lexical category, it has to have a morphophonological structure that can plausibly be reanalysed
as an inflected form (Willis 2007: 303). Evidently, this becomes decreasingly likely the more
Nevertheless, Norde (2009: 135ff.) lists a number of cases. These are: the preterite
subjunctive of modal welle ‘to want to’ > full verb wotte ‘to wish’, in a variety of Pennsylvania
German spoken in Waterloo County, Canada (Burridge 1998); the Chinese deontic modal dĕi
> lexical verb meaning ‘to need, require’ (Ziegeler 2004); Old Church Slavonic nĕčĭto
‘something’ > Bulgarian nešto ‘thing’ (Willis 2007: 278ff.), or Old Irish ní ‘something’ >
Modern Irish ní ‘thing’ (Willis 2016: 205); the Welsh 3SG possessive pronoun eiddo ‘his’ >
noun meaning ‘property’ (Willis 2007:283ff.); and the Middle Welsh preposition yn ôl ‘after’
> full verb nôl ‘to fetch (Willis 2007: 292ff., Trousdale & Norde 2013: 38f.).
ôl. In late Middle Welsh, yn ol is attested in constructions such as (9), where it is ambiguous
between ‘after’ and ‘fetch’. Subsequently a formal split occurs, with yn ol continuing as a
preposition while the verb is reduced to nol (Modern Welsh nôl). This split was facilitated by a
reanalysis from yn ôl as y nol, that is, a purposive marker y plus a verb. From the 17th century
onwards, examples featuring verbal inflections, such as (10), begin to appear (Willis 2007: 294,
297).
then PRT went the lads after his horse and his weapons for Arthur
‘Then the lads went after / went to fetch his horse and his weapons for Arthur’
inflections. Similar effects are seen in other cases of degrammation. Thus, the Pennsylvania
German lexical verb wotte ‘to wish’, deriving from an indeclinable modal form ‘would’,
features inflected forms such as imperative wott or the past participle gewott. As an example
from the nominal domain, the Bulgarian noun nešto ‘thing’ (from a pronoun meaning
‘something’) has regular neuter inflection (plural nešta, definite form neštoto, with enclitic
definite article).
3.5 Deinflectionalization: decreasing morphophonological integration
because it involves a shift from an inflectional affix to another type of bound morpheme (e.g. a
derivational affix or a clitic). This naturally raises the question of why these shifts should be
considered degrammaticalization and not, for instance, lateral conversions on the level of bound
morphemes. A full discussion of this issue falls outside of the scope of this contribution (for
this, see Norde 2009: 152ff.), but some brief observations are in order here. First, what
distinguishes inflectional affixes from derivational ones is that the former are grammatically
obligatory, whereas the latter are not (Bybee 1985: 81ff.). For instance, case affixes are required
by a specific morpho-syntactic context, but derivational affixes add lexical content (sometimes
changing word class). Furthermore, the scope of inflectional affixes is usually restricted to a
single stem, whereas derivational affixes may take scope over an entire phrase (see Norde &
Van Goethem 2015 for examples from Dutch). On the basis of this, it could be argued that this
makes inflectional affixes more strongly integrated in the grammar of a language, and hence
‘more grammatical’. As far as the differences between inflectional affixes and clitics are
concerned (see e.g. Zwicky & Pullum 1983), it can be noted that inflectional affixes are often
restricted to a specific inflectional class, while clitics are far less selective with respect to the
host they attach to. Also, morphophonological idiosyncrasies are more commonly found in
inflection, for instance when the affix changes vowel quality or quantity of the root. The lesser
degrammaticalized status. This is also in line with the cline in (1), where inflectional affixes are
closer to the grammaticalized pole than are clitics, as well as with the fact that a change from
MASC.SG.NOM –er > Modern Swedish derivational nominalization suffix –er (Norde &
Trousdale 2016: 181ff.); Old Swedish NEUT.PL.NOM/ACC –on > Modern Swedish derivational
‘berry-suffix’ –on (Norde 2009: 181f.); and Kwazavii inflectional exhortative marker –ni >
The latter case is illustrated in (11) and (12). The primary reasons for analysing the
suffix -nĩ as derivational are the following (Van der Voort 2002:312ff. and personal
communication): causational –nĩ does not belong to a paradigm and is optional, whereas
exhortative –ni forms part of a verbal mood paradigm and is grammatically obligatory.
Furthermore, –nĩ can receive primary stress, which in Kwaza is always realized on the final
already drink-3-EXH-1SG-DEC
already drink-CAUS-1SG-DEC
Examples of deinflectionalization from affix to clitic are Russian 2PL imperative suffix -t’e >
enclitic =t’e, with palatalization of the consonant before [t’] retained (suffixation would have
resulted in depalatalization), which later spread to 1PL as a hortative clitic (Andersen 2008: 28);
1894; Allen 2008), Danish (Herslund 2001; Perridon 2013), Norwegian (Johannessen 1989),
and Swedish (Norde, 1997, 2006a, 2013; Trousdale & Norde 2013: 40ff.; Norde & Trousdale
2016: 172 ff.; Piotrowska 2017). A recent example is discussed in Haig (2018) and concerns
the deinflectionalization of agreement markers in Central Kurdish past tense transitive verbs.
In this language, subject agreement markers are obligatory in the past tense, whether or not an
overt pronoun (e.g. eto in (13)) is present. Object agreement markers, on the other hand, are no
longer obligatory. They appear when there is no free pronoun present, as in (14), but are
excluded when there is a free pronoun, as in (15) (Haig 2018: 803). In other words, free
pronouns and bound object agreement markers have become alternating person indices.
‘You left.’
otherwise IPFV=3s:A=kill.PST-3PL:P
The bound markers furthermore show signs of decreasing phonological integration, e.g.
because they can be separated from their stem by an intervening subject-indexing clitic, as
(16) nārd=mān-in
send.PST=1PL.A-3PL.P
these agreement markers cannot be affixes, since affixes do not normally attach to material
Unlike the previous two types of degrammaticalization, debonding is relatively common, but it
is also less homogeneous. This is because its input can be inflectional affixes, derivational
affixes, clitics, or affixoids. In debonding, the major change is on the level of morphology – a
bound morpheme becomes free with only minimal changes to its semantics, or indeed none at
all. Examples of debonded inflectional affixes include Irish muid (1PL. verb suffix > personal
pronoun ‘we’; Doyle 2002: 68) and North Saami haga ‘without’ Ylikoski 2016), , exemplified
in (17). The North Saami postposition haga in (17a) derives from a Proto-Saami bound abessive
case form *-ptāke̮k or *-ptāke̮n, that is still found as an abessive suffix in other Saami languages,
Another interesting example from North Saami, likewise discussed by Ylikoski, concerns the
suffix –naga ‘stained with’ (possibly based on the essive case suffix –na, see Ylikoski 2016 for
details). This morpheme comes in several forms (Ylikoski 2016: 130, 141, 142): as a suffix in
for plural it is clearly a noun, not a postposition. Since this is a gradual development, naga
would be the first known example of a degrammaticalization chain, where all other
Debonding of clitics is attested, for instance, in English (Fischer 2000; Fitzmaurice 2000) and
Norwegian (Faarlund 2007a) infinitive markers, which used to be proclitic. The change is
grammatical function they had as bound morphemes. Derivational affixes, on the other hand,
do not have a grammatical function, so when these debond we see an increase in semantic
substance. For example, the Dutch numeral suffix –tig only has modifying semantics ‘multiply
Similar observations can be made for other debonding derivational affixes, such as English -ish
(Kuzmack 2007; Norde 2009: 223ff.; Traugott & Trousdale 2013: 233ff.; Bochnak & Csipak
2014; Pierce 2014, 2015), or Turaix –lá ‘(approx.) anywhere’ > separable derivational marker
with the additional function of focalizer (Idiatov 2008; Norde 2009: 225ff.).
The final morpheme type to be discussed in this section are affixoids, “which look like
parts of compounds, and do occur as lexemes, but have a specific and more restricted meaning
when used as part of a compound” (Booij 2009: 208; see also Booij 2017). Affixoids are found
in all Germanic languages (if only marginally in English), and can combine with both nouns
and adjectives / adverbs. Some examples of bound adjectival affixoid constructions, in which
the affixoid has intensifying function, are given in (23). Free affixoids are illustrated in (24)
Debonding of affixoids has been the topic of a large number of recent publications and
references therein, among them Norde & Van Goethem 2014, 2015, 2018; Van Goethem &
Hiligsmann 2014; Van Goethem & De Smet 2014; Van Goethem and Hüning 2015; and
Battefeld, Leuschner & Rawoens 2018. While some free affixoids are merely orthographic
variants of the bound form (Norde & Van Goethem 2014 makes this case for Swedish jätte),
many others can be considered different constructions, primarily in terms of productivity and
collocational properties. For example, Dutch kei (originally ‘boulder’) is primarily used in the
simile construction keihard ‘hard as a boulder, very hard’, whereas free kei primarily collocates
4.1 Introduction
Grammaticalization and degrammaticalization are not the only types of composite change that
involve primitive changes at more than one level, including morphology. x The most obvious
candidate is lexicalization (Brinton & Traugott 2005), discussed in the next three sections of
this chapter. Other types of change include category change without (de)grammaticalization
and exaptation. The lines between the change types in this section and either grammaticalization
or degrammaticalization are not clear-cut, and therefore some authors have treated them as
Derivational affixes are bound morphemes that can be traced back to various sources, primarily
phrases and compounds. An example of a derivational suffix emerging out of a phrase, is the
French adverbial suffix –ment, which ultimately derives from the ablative singular of the Latin
noun mens (GEN mentis) ‘mind’, e.g. Latin placida mente ‘with a quiet mind’ > French
placidement ‘quietly’.xi Suffixes deriving from compounds include English –hood (< Old
English had ‘rank, character, state’ (Brinton & Traugott 2005: 98), or German –schaft (e.g. in
Gastgeberschaft ‘hosting’) < Old High German scaf ‘creature, quality’ (Nübling et al. 2017:
98).
of grammaticalization, the question of whether derivational affixes, too, can be considered the
result of grammaticalization is less easy to answer. Although derivational suffixes show some
independent morpheme status), they do not have a grammatical function, but a lexical one: to
derive new lexemes. Some scholars have therefore suggested that the emergence of derivational
affixes may be a process sui generis (e.g. Himmelmann 2004), whereas others (e.g. Lightfoot
2005), point out that derivational affixes exhibit properties of both grammaticalization and
lexicalization (for more extensive discussion of these issues see Brinton & Traugott 2005, or
Norde 2009: 11ff.). Following these authors, derivational affixes were not discussed in the
distinguish it from other types of lexicalization (for a recent survey see Hilpert 2019). This
change concerns bound morphemes that come to be employed as nouns or verbs, e.g. when ism
is used as a hypernym for ‘ideology’ that can be pluralized (English isms, Dutch ismen, Swedish
ismer, French ismes, etc.). The autonomous usage of bound morphemes is easily confused with
degrammaticalization (more specifically, debonding), and indeed some authors have interpreted
Newmeyer 1998:269f.; Brinton & Traugott 2005:60; see Van der Auwera 2002 for discussion).
Hopper & Traugott (2003: 134), on the other hand, characterize cases such as isms as “the
recruitment of linguistic material to enrich the lexicon.” What crucially sets isms and similar
developments apart from debonding is that their ‘constructional identity’ (see 2.1) is not
preserved – there are no bridging contexts in which ism can be interpreted as either a suffix or
a noun. Furthermore, ism can be said to have ‘jumped the cline’ from affix to major category,
Although (de)grammaticalization often involves a shift from one category to the other, the
reverse is not necessarily true. Lexical items may change category without
words and affixes (Norde 2009: 112ff.). Lateral shifts are conversions from one major category
to another, for instance by means of adding word-class changing derivational suffixes, e.g.
English beauty (N) > beautiful (ADJ), or swim (N) > swimmer (V). Other lateral shifts include
zero derivations, e.g. English text (N) > to text (V); reduplication, e.g. Bahasa Indonesia keras
‘loud’ > keras-keras ‘loudly’ (Sneddon 1996: 19); or templatic derivation, e.g. Arabic k-t-b ‘to
write’ > kaatib ‘writer’ (Benmamoun 2016: 65).xii Alternatively, a word may shift category
adjective by adding a comparative or superlative suffix, as in the Dutch example in (26) (Booij
& Audring 2018: 216) xiii and the Italian example in (27) (Grandi, Nissim & Tamburini 2011:
174) respectively.
Lexicalization of function words is attested when minor categories (function words) can be
converted into major categories, as in English to off, to down (adverb > verb), or ifs and buts
(conjunction > Noun) (Brinton & Traugott 2005: 38). Similarly, several languages can verbalize
a 2SG pronoun by adding verbal suffixes, as in French tutoyer (< tu, toi), Swedish dua (< du),
or Dutch jijen en jouen (< jij, jou) ‘to use the informal pronoun of address’ (Norde 2009: 113).
4.5 Exaptation
(see 1990 and 1997: 316ff.), to refer to the refunctionalization of an (often bound) morpheme
(for a recent survey see Norde & Van de Velde 2016). This functional ‘leap’ distinguishes
2016). An example of exaptation is the rise of the plural marker –er in German, which derives
from a stem-forming suffix –ir that was originally restricted to a small set of 10 neuter nouns,
mostly denoting small livestock (e.g. lamm ‘lamb’ or huon ‘chicken’). It was already
desemanticized in Old High German, and with the stem-forming suffix disappearing from the
singular as a result of reductive phonological changes, -ir was reanalysed as a plural marker.
Once that had happened, -er spread to other declensions and is found in approximately 100
.
Early OHG Late OHG MoG
Singular Nominative lamb lamb Lamm
Genitive lemb-ir-es lamb-es Lamm-es
Dative lemb-ir-e lamb-e Lamm
Accusative lamb lamb Lamm
Plural Nominative lemb-ir lemb-ir Lämm-er
Genitive lemb-ir-o lemb-ir-o Lämm-er
Dative lemb-ir-um lemb-ir-um Lämm-er-n
Accusative lemb-ir lemb-ir Lämm-er
Table 3: Exaptation of Old High German –ir (Szczepaniak 2011: 57)xv
In the case of Old High German –ir, then, exaptation leads to a ‘reorganization’ of an existing
paradigm, rather than to a new one. This distinguishes exaptation from secondary
grammaticalization (e.g. the Romance inflectional future discussed in 3.3). Furthermore, there
either.xvi
(De)grammaticalization and morphology are linked to each other in myriad ways. Inflections
may be lost (in primary grammaticalization) or acquired (in degrammation). New inflections,
or indeed entire inflectional categories, may grow out of secondary grammaticalization. Bound
deinflectionalization) or even start a new life as free morphemes (debonding). Whatever the
type of (de)grammaticalization involved, these are all gradual changes that take place in specific
constructional contexts, and they are typically accompanied by changes on other levels
(phonology, semantics, pragmatics, and / or syntax). Naturally, this does not imply that the
Morphological change as part of composite changes is, in other words, a research topic of vast
proportions, and only a handful examples could be quoted in this contribution. In spite of the
vast literature on the topic, however, the interplay between morphology and other levels,
References
Allen, Cynthia L. 2008. Genitives in Early English. Oxford: Oxford University Press.
Eythórsson, Thórhallur (Ed.) Grammatical change and linguistic theory. The Rosendal
Anttila, Raimo. 1989. Historical and comparative linguistics. Amsterdam / Philadelphia: John
Benjamins.
Barðdal, Jóhanna, Elena Smirnova, Lotte Sommerer & Spike Gildea (Eds.). 2015. Diachronic
German, Dutch and Swedish. Constructional networks and the loci of change. In Van
Goethem, Kristel, Muriel Norde, Evie Coussé & Gudrun Vanderbauwhede (Eds.) Category
Benjamins.
Benmamoun, Elabbas. 2016. Verbal and nominal plurals and the syntaxmorphology interface.
In Davis, Stuart & Usama Soltan (Eds.) Perspectives on Arabic Linguistics XXVII, 59-74.
Bergs, Alexander, & Gabriele Diewald (Eds.). 2008. Constructions and Language Change.
https://oxfordre.com/linguistics/view/10.1093/acrefore/9780199384655.001.0001/acrefore-
9780199384655-e-103.
Bisang, Walter, Nikolaus P. Himmelmann & Björn Wiemer (Eds.). 2004. What Makes
Grammaticalization? A look from its fringes and its components. Berlin & New York:
Mouton de Gruyter.
Bochnak, M. Ryan & Eva Csipak. 2014. A new metalinguistic degree morpheme. In Todd
Snider, Sarah D’Antonio and Mia Weigand (Eds.) Proceedings of Semantics and Linguistic
Booij, Geert. 2009. Compounding and Construction Morphology. In Lieber, Rochelle & Pavol
University Press.
9780199384655-e-254.
Booij, Geert & Jenny Audring. 2018. Category change in construction morphology. In Van
Goethem, Kristel, Muriel Norde, Evie Coussé & Gudrun Vanderbauwhede (Eds.) Category
Benjamins.
Börjars, Kersti & Pauline Harries. 2008. The clitic-affix distinction, historical change, and
Breban, Tine & Svenja Kranich (Eds.). 2015. What happens after grammaticalization?
Secondary grammaticalization and other late stage processes. Language Sciences 47:B.
Brinton, Laurel J. & Elizabeth Closs Traugott. 2005. Lexicalization and language change.
Burridge, Kate. 1998. From modal auxiliary to lexical verb. The curious case of Pennsylvania
German wotte. In Hogg, Richard M. & Linda van Bergen (Eds.) Historical linguistics 1995.
Bybee, Joan. 1985. Morphology. A study of the relation between meaning and form. Amsterdam
Bybee, Joan. 2007. Frequency of use and the organization of language. Oxford: Oxford
University Press.
Bybee, Joan. 2010. Language, usage and cognition. Cambridge: Cambridge University Press.
Hoffmann, Thomas & Graeme Trousdale (Eds.) The Oxford handbook of construction
aspect and modality in the languages of the world. Chicago & London: The University of
Chicago Press.
23: 2-3.Coussé, Evie, Peter Andersson & Joel Olofsson. 2018. Grammaticalization meets
Coussé, Evie, Peter Andersson & Joel Olofsson (Eds.) Grammaticalization meets
approach with particular reference to Ancient Greek. In Giacalone Ramat, Anna & Paul J.
Benjamins.
Croft, William. 2001. Radical Construction Grammar. Oxford: Oxford University Press.
Davidse, Kristin, Tine Breban, Lot Brems & Tanja Mortelmans (Eds.) 2012.
John Benjamins.
Davis, Stuart & Natsuko Tsujimura. 2014. Non-Concatenative Derivation. Other Processes. In
Lieber, Rochelle & Pavol Štekauer (Eds.), The Oxford Handbook of Derivational
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199641642.001.0001/oxfo
rdhb-9780199641642-e-012.
De Clerck, Bernard & Timothy Colleman. 2013. From noun to intensifier. Massa and massa’s
De Lancey, Scott. 1994. Grammaticalization and linguistic theory. In Gomez de Garcia, Jule &
David S. Rood (Eds.) Proceedings of the 1993 Mid-America linguistics conference and
Detges, Ulrich. 1998. Echt die Wahrheit sagen. Überlegungen zur Grammatikalisierung von
Detges, Ulrich & Richard Waltereit. 2002. Grammaticalization vs. Reanalysis. A semantic-
151-195.
Faarlund, Jan Terje. 2007b. From clitic to affix. The Norwegian definite article. Working papers
the infinitive in English. In Fischer, Olga, Anette Rosenbach & Dieter Stein (eds) Pathways
Benjamins.
Fischer, Olga, Muriel Norde & Harry Perridon (Eds.). 2004. Up and down the cline. The nature
American English. In Fischer, Olga, Anette Rosenbach & Dieter Stein (eds) Pathways of
Benjamins.
Fleischman, Suzanne. 1982. The future in thought and language. Diachronic evidence from
Ramat, Anna & Paul J. Hopper (Eds.) The limits of grammaticalization, 107-127.
Giacalone Ramat, Anna & Paul J. Hopper (Eds.). 1998. The limits of grammaticalization.
Givón, Talmy. 1991. The evolution of dependent clause morpho-syntax in Biblical Hebrew. In
Traugott, Elizabeth Closs & Bernd Heine (Eds.) Approaches to grammaticalization vol. II,
Grandi, Nicola, Malvina Nissim & Fabio Tamburini. 2011. Noun-clad adjectives. On the
Haspelmath, Martin. 1993. The diachronic externalization of inflection. Linguistics 31:2, 279-
309.
1068.
grammaticalization. In Fischer, Olga, Muriel Norde & Harry Perridon (eds) Up and down
Benjamins.
Heine, Bernd. 1997. Cognitive foundations of grammar. Oxford: Oxford University Press.
Heine, Bernd. 2002. On the role of context in grammaticalization. In Wischer, Ilse & Gabriela
John Benjamins.
Heine, Bernd. 2011. Grammaticalization in African languages. In Narrog, Heiko & Bernd
University Press.
Heine, Bernd, Heiko Narrog & Haiping Long. 2016. Constructional change vs.
Heine, Bernd & Tania Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge:
Herslund, Michael. 2001. The Danish s-genitive. From affix to clitic. Acta Linguistica
https://oxfordre.com/linguistics/view/10.1093/acrefore/9780199384655.001.0001/acrefore-
9780199384655-e-622.
orthogonal? In Bisang, Walter, Nikolaus P. Himmelmann & Björn Wiemer (eds) What
makes grammaticalization? A look from its fringes and components, 21-42. Berlin / New
Hoeningswald, Henry M. 1966. Are there universals of linguistic change? In Greenberg, Joseph
Hopper, Paul. J. & Elizabeth Closs Traugott. 2003. Grammaticalization. Second edition.
Seoane, Elena, María José López-Couso, in collaboration with Teresa Fanego (Eds.)
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199641642.001.0001/oxfo
rdhb-9780199641642-e-011.
340.
Johannesen, Janne Bondi, 1989. Klitika — en avgrensning. Norsk lingvistisk tidsskrift 1989:2,
117-147.
Joseph, Brian D. 2001. Is there such a thing as grammaticalization? Language Sciences 23:2-3,
163-186.
Larm, Karl. 1936. Den bestämda artikeln i äldre fornsvenska. En historisk-syntaktisk studie.
Lass, Roger 1990. How to do things with junk. Exaptation in language evolution. Journal of
University Press.
Lincom Europa.
http://oxfordre.com/linguistics/view/10.1093/acrefore/9780199384655.001.0001/acrefore-
9780199384655-e-248.
López-Couso, María José & Elena Seoane (Eds.). 2008. Rethinking grammaticalization. New
Lorenz, David. 2013. From reduction to emancipation: Is gonna a word? In Hasselgård, Hilde,
Jarle Ebeling and Signe Oksefjell Ebeling (Eds.) Corpus Perspectives on Patterns of Lexis,
Narrog, Heiko. 2016. Exaptation in Japanese and beyond. In Norde, Muriel & Freek Van de
Velde (Eds.) Exaptation and language change, 93-120. Amsterdam / Philadelphia: John
Benjamins.
Newmeyer, Frederick J. 1998. Language Form and Language Function. Cambrige MA: MIT
Press.
Noël, Dirk. 2007. Diachronic construction grammar and grammaticalization theory. Functions
Norde, Muriel. 2006b. Van suffix tot telwoord tot bijwoord: degrammaticalisering en
Norde, Muriel. 2011. Degrammaticalization. In Narrog, Heiko & Bernd Heine (Eds.) The
Norde, Muriel. 2012. Lehmann’s Parameters Revisited. In Davidse, Kristin, Tine Breban, Lot
Brems & Tanja Mortelmans (Eds.) Grammaticalization and Language Change: New
Norde, Muriel. 2013. Tracing the Origins of the Swedish Group Genitive. In Carlier, Anne &
Jean-Christophe Verstraete (Eds.) The Genitive (= Case and Grammatical Relations across
Norde, Muriel & Karin Beijering. 2014. Facing interfaces. A clustering approach to
Norde, Muriel, Karin Beijering & Alexandra Lenz (Eds.). 2013. Current trends in
Norde, Muriel & Graeme Trousdale. 2016. Exaptation from the perspective of construction
morphology. In Norde, Muriel & Freek Van de Velde (Eds.), Exaptation and language
Norde, Muriel & Kristel Van Goethem. 2014. Bleaching, productivity and debonding of
Norde, Muriel & Kristel Van Goethem. 2018. Debonding and clipping of prefixoids in
Springer.
Norde, Muriel & Freek Van de Velde (Eds.). 2016. Exaptation and language change.
Nübling, Damaris, Antje Dammel, Janet Duke & Renata Szczepaniak. 2017. Historische
Perridon, Harry. 2013. The emergence of the s-genitive in Danish. Language Sciences 36, 134-
146.
Pierce, Marc. 2014. The further degrammaticalization of ish. American speech 89:1, 115-118.
Pinkster, Harm. 1987. The strategy and chronology of the development of future and perfect
tense auxiliaries in Latin. In Harris, Martin & Paolo Ramat (eds) Historical development of
Piotrowska, Alicja. 2017. Mannen på gatans åsikt och personen bredvids förvåning – s-genitiv
John Rennison & Friedrich Neubarth (eds) Naturally! Linguistic studies in honour of
Seoane, Elena & María José López-Couso (Eds.). 2008. Theoretical and empirical issues in
Smirnova, Elena. 2015. When secondary grammaticalization starts. A look from the
Smith, Andrew D. M., Graeme Trousdale & Richard Waltereit (Eds.). 2015. New directions in
Sneddon, James Neil. 1996. Indonesian. A comprehensive grammar. London / New York:
Routledge.
Stroh-Wollin, Ulla. 2015. Understanding the gradual development of definiteness marking: the
Traugott, Elizabeth Closs. 2002. From etymology to historical pragmatics. In Minkova, Donka
& Robert Stockwell (Eds.) Studies in the history of the English language, 19-49. Berlin /
Traugott, Elizabeth Closs. 2007. The concepts of constructional mismatch and type-shifting
In Eckardt, Regine, Gerhard Jäger & Tonjes Veenstra (Eds.) Variation, Selection,
Development. Probing the Evolutionary Model of Language Change, 219-250. Berlin / New
change. In Hancil, Sylvie & Ekkehard König (Eds.) Grammaticalization – Theory and Data,
Traugott, Elizabeth Closs & Bernd Heine. 1991a (Eds.). Approaches to Grammaticalization,
Traugott, Elizabeth Closs & Bernd Heine. 1991b (Eds.). Approaches to Grammaticalization,
Traugott, Elizabeth Closs & Bernd Heine. 1991c. Introduction. In Traugott, Elizabeth C. &
Traugott, Elizabeth Closs & Graeme Trousdale. 2010. Gradience, gradualness and
18 Jun. 2019.
https://oxfordre.com/linguistics/view/10.1093/acrefore/9780199384655.001.0001/acrefore-
9780199384655-e-260.
from the history of a composite predicate construction in English. In Trousdale, Graeme &
constructions. In Davidse, Kristin, Tine Breban, Lot Brems & Tanja Mortelmans (Eds.)
Grammaticalization and Language Change: New reflections, 167-198. Amsterdam &
Van der Auwera, Johan. 2002. More Thoughts on Degrammaticalization. In Wischer, Ilse &
Van der Voort, Hein. 2002. The quotative construction in Kwaza and its (de-
)grammaticalisation. In Crevels, Mily, Simon van de Kerke, Sérgio Meira & Hein van der
Voort (Eds.) Selected papers from the 50th international congress of Americanists in Warsaw
and the Spinoza workshop on Amerindian languages in Leiden, 307-328. Leiden: Research
Van der Voort, Hein. 2004. A grammar of Kwaza. Berlin / New York: Mouton de Gruyter.
Van Goethem, Kristel & Hendrik De Smet. 2014. How nouns turn into adjectives. The
emergence of new adjectives in French, English and Dutch through debonding processes.
Van Goethem, Kristel., & Hiligsmann, Philippe. 2014. When two paths converge. Debonding
and clipping of Dutch reuze ‘giant; great’. Journal of Germanic Linguistics 26:1, 31-64.
Van Goethem, Kristel., & Hüning, Matthias. 2015. From noun to evaluative adjective:
Conversion or debonding? Dutch top and its equivalents in German. Journal of Germanic
Van Pareren, Remco. 2013. Body part terms as a semantic basis for grammaticalization: a
Mordvin case study into spatial reference and beyond. Language Sciences 36, 90-102.
Von Mengden, Ferdinand. 2016. Functional changes and (meta-)linguistic evolution. In Norde,
Muriel & Freek Van de Velde (Eds.) Exaptation and language change, 121-162. Amsterdam
Von Mengden, Ferdinand & Horst Simon. 2014. Refining grammaticalization. Folia
Linguistica 48:2.
45:2, 271-310.
of abductive reanalysis. In Norde, Muriel & Freek Van de Velde (eds) Exaptation and
Wischer, Ilse & Gabriela Diewald (Eds.). 2002. New reflections on grammaticalization.
adverbs and a free lexical noun from inflectional and derivational suffixes. Finnisch-
Ziegeler, Debra. 2004. Redefining unidirectionality. Is there life after modality? In Fischer,
Olga, Muriel Norde & Harry Perridon (eds) Up and down the cline the nature of
Zwicky, Arnold M. & G. K. Pullum. 1983. Cliticization vs inflection: English n’t. Language
59:3,502-51.
Notes
i
For a careful account of the development of the framework since Meillet’s publication see Von Mengden 2016.
ii
This is essentially an empirical observation, but chances that the same gram will go through a series of
degrammaticalization changes are extremely low for reasons that are not in the scope of the present paper (see
Norde 2009: 100ff. for extensive discussion ).
iii
Note that Norde’s definition of ‘degrammation’ is slightly different from Andersen’s (2006: 232), which refers
to the loss of grammatical content. This includes, e.g., the loss of grammatical content in univerbation, as in
English tomorrow.
iv
In languages where minor categories may be inflected as well, primary grammaticalization does not
necessarily lead to the loss of inflections (although there may be different inflections). One example are Mordvin
(inflected) adpositions that grammaticalized from body part nouns (Van Pareren 2013).
v
In its cliticized form, =(h)inn was often ambiguous between demonstrative and determiner uses, but this issue
cannot be discussed here. In the Modern Swedish example, the s-genitive is marked with a clitic symbol (=),
because it has degrammaticalized from affix to clitic (Norde 2006a, 2009: 160ff.).
vi
Haspelmath does not mention an intermediary clitic change, but it is unclear whether it was never there, or
whether no evidence of a clitic survives.
vii
Kwaza is an isolated language spoken in Rondônia, Brazil (Van der Voort 2004).
viii
Source: http://www.degoednieuwskrant.nl/index.php/nieuws/gezondheid/1913-tig-redenen-om-chocolade-te-
eten.
ix
Tura belongs to the Eastern Mande languages, a subbranch of Niger-Congo.
x
For morphological change and its causes in general see Trips 2017.
xi
The suffix has cognates in all Romance languages except Romanian, cf. Portuguese cruamente ‘cruelly’,
Spanish distintamente ‘distinctly’, Italian raramente ‘rarely’, Occitan and Catalan bellamen(t) ‘beautifully’, and
Sardinian finalmenti(s) ‘finally’ ((Hoenigswald 1966: 44; Hopper & Traugott 2003:141f.; Lehmann 1995:87,
Giacalone Ramat 1998: 120; Detges 1998, Torner 2005, Norde 2009: 41ff.).
xii
For additional examples and surveys of the literature see Davis & Tsujimura 2014, Inkelas 2014 or Lieber
2017.
xiii
The Dutch examples if from the poem ‘Grootouders’ by Judith Herzberg, in Soms vaak (2004).
xiv
Source: www.tvblog.it.
xv
The stem vowel change is due to i-mutation.
xvi
There are a few known examples where a suffix deinflectionalizes while also being exapted for a new
function, but this appears to be very rare (Norde & Trousdale 2016).