1 s2.0 S004452311500008X Main

Zoologischer Anzeiger 256 (2015) 42–53
Contents lists available at ScienceDirect
Zoologischer Anzeiger
journal homepage: www.elsevier.com/locate/jcz
Word forms, classification, and family trees of languages—Why

morphology is crucial for linguistics
Lars Erik Zeige ∗
Humboldt University, Berlin, Germany
a r t i c l e i n f o a b s t r a c t
Article history: ‘Morphology’ in linguistics is the study of the structure and function of word forms. In this paper, Sec-
Received 2 October 2014 tions “the structure of word forms: elements and relations” and “the function of word forms: values and
Received in revised form 19 February 2015 systems” will give an insight into the basic notions and subfields of linguistic morphology to illustrate
Accepted 19 February 2015
the linguistic approach to structure and function. It will then proceed to identify the position of morphol-
Available online 21 February 2015
ogy within linguistics and the repeated conjunctions between biology and linguistics by glancing at the
theoretical foundations (Section “foundations”) and the history (Section “contributions of morphology
Keywords:
to the language sciences”) of morphology in linguistics as well as today’s theoretical and methodolog-
Morphology
Linguistics
ical challenges (Section “modern challenges”). The paper will conclude with some deliberations on the
Form relevance of morphological studies as part of the academic canon.
Classification © 2015 Elsevier GmbH. All rights reserved.
Family tree models
Phylogenetic models
Historical-comparative linguistics
Preface itself can be thought of as the abstracted inventory of these

lexemes, the result of linguistic research much in the sense of a
Amō, amās, amat, amāmus, amātis, amant—from your time at lexicon in a book shelf, or as the neuro-cognitive storage facility
school, you may remember, with pleasure or dismay, swotting in human brains. Take for example, the lexeme walk (the notation
paradigms of words from a foreign language. In the middle ages, in capitals following Lyons, 1968) and its word forms walk, walks,
Latin regula ‘rule’ referred to both the formation rules producing walked, walking. These word forms are the written and spoken
the word forms and the paradigm listing the word forms as a whole. forms which represent the lexeme in speech. Morphology studies
The paradigm, in turn, was the basis on which the rules would be the structure and function of these word forms. Just by glancing
learnt and internalized (Law, 2003: 83–85). Regula meant origi- at the example of walk, walks, walked, walking, it is apparent that
nally ‘bar, batten, measuring stick’ and is etymologically related word forms exhibit structural properties which rest on a sub-word
to Modern English rule and ruler—with the latter sometimes being level. It is therefore, necessary to have concepts and a terminology
instrumental in the learning of the former. But as a matter of fact, available which allow identifying the inventory of constructive
you tacitly acquired a great deal of abstract morphological knowl- entities and the rules of combination which constitute word forms.
edge by repeating paradigms in a particular fashion. Regarding the inventory of constructive entities, the linchpin
of modern morphological study is the notion of the ‘morpheme’,
1. The structure of word forms: elements and relations established during the 1880s and first defined in 1895 by Jan
Baudouin de Courtenay (de Courtenay, 1972 [1895]: 153), later
‘Morphology’ in linguistics is the study of the structure and canonically defined by Hockett (1958: 264): a morpheme is the
function of word forms. It is not, however, the study of ‘words’, smallest, indivisible unit of grammatical function and form. It can be
understood as the elementary units of an utterance possessing ‘bound’ or ‘free’. A free morpheme can function independently, like
semantic or pragmatic meaning. In linguistics, ‘word’ is generally car in I bought a new car. A bound morpheme depends in its actual-
abandoned in favor of two more specific terms: ‘lexeme’ and ization on at least one free morpheme, like {–ing} as part of walking
‘word form’. Following Matthews (1974: 26), the lexeme is the in He is walking down Baker Street. Morphemes can also be classified
fundamental unit found in the ‘lexicon’ of a language. The lexicon as ‘lexical’ or ‘grammatical’. Lexical morphemes possess a notional
meaning—for example, individuals, objects or situational notions,
both real or imagined. Grammatical morphemes, in contrast, pos-
∗ Tel.: +49 3020939677; fax: +49 3020939703. sess a meaning that is best expressed in terms of grammatical
http://dx.doi.org/10.1016/j.jcz.2015.02.003
0044-5231/© 2015 Elsevier GmbH. All rights reserved.
L.E. Zeige / Zoologischer Anzeiger 256 (2015) 42–53 43
categories, such as {–s} in (she) walks carries the abstracted mean- word class of the derivative. For example, the derivational suf-
ing of third person singular and present tense indicative active fix {–ness} usually attaches to adjectives to form a noun. There
(walk-3sg.Pres.Ind.Act). The classifications free/bound and lexi- is no *walkness (verb) or *bookness (noun). The derivational suf-
cal/grammatical overlap: although free morphemes for the most fix {–hood}, by contrast, usually combines with a noun to form
part are also lexical (cat, nice, well, sing), there are also free gram- a noun, as in childhood. It does not, however, combine with
matical morphemes (with, he, the, because). In turn, there are also adjectives (*coldhood) or verbs (*walkhood). Derivated word forms
bound lexical morphemes, which do not exist alone (cf. *huckle in can also serve as stems for further affixation (to be-head some-
huckleberry). body > the behead-ing of somebody). Derivation, in general, is about
Regarding the rules which govern the way morphemes are com- the morphological restrictions of combining lexical and grammat-
bined, morphology analyses word forms by three main areas of ical morphemes.
interest—inflection, derivation, and compounding. Traditionally, In compounding, two lexical morphemes attach. For example, in
these are thought to be dichotomous. Inflection, on the one hand, windmill a noun and another noun combine to form a noun, in pick-
creates different word forms of a given lexeme (walk-ing, walk-s, pocket a verb attaches to a noun to form a noun, in bowhunt a noun
walk-ed), while derivation and compounding, on the other hand, and verb combine to form a verb (its gerund producing the noun
constitute means of word formation, the creation of new lexemes bowhunting), or in deepfry an adjective attaches to a verb to form a
(speech-less, speech analysis). verb (its past participle giving us the adjective deepfried). Again,
As an example for inflectional morphology, let us expand the compound words can serve as a base for further compounding,
paradigm of the Latin verb amō ‘to love’. most famously in German, giving rise to urban legends, such as the
Donaudampfschiffahrtskapitänswitwenrente (‘pension for widows of
amō amābam amābor
amās amābās amāberis
captains of steam boats on the Danube’). In general, compound-
amat amābat amābitur ing is about the morphological restrictions of combining lexical
amāmus amābāmus amābimur morphemes with lexical morphemes.
amātis amābātis amābiminı̄ On the whole, to morphologically define parts of speech works
amant amābant amābuntur
with the necessary precision, particularly in the case of verbs,
The paradigm of a verb assembles in a systematic fashion the nouns, and adjectives. But as is the nature of all classifications,
1st, 2nd, and 3rd person in both singular and plural number, as well these definitions have overlaps and areas of transition, and, as with
as the tenses (here e.g., present, imperfect, and future), the active all progress in sciences, the main interest of research is now put
and passive voice (column one and two vs. column three), and the onto the fringes of categories and the irregularities of its objects.
grammatical mood (here only indicative; not shown subjunctive, One source of confusion is the difference between the word class
imperative). The inflected forms contain a lexical morpheme as a of the lexeme and the syntactic function a word form can fulfill.
root and a grammatical morpheme carrying the inflectional infor- For example, an adjective can take the position of an adverb (Still
mation. water runs deep/His contribution is still due), or a preposition can
Apart from giving an overview of all the inflected forms of a function as a subordinating conjunction (I have been waiting since
verb and the possibility of learning, paradigms allow for a very five o’clock/Since there is no other option, the villagers accept their
substantial contribution to understanding the inner structure of fate). One solution is to consider them as separate words of equal
languages: when collecting all possible inflected word forms in form belonging to different classes. Another source of disorder is
paradigms, it becomes apparent that some words fill the boxes e.g., words that either unite the properties of several well established
number and tense, while others do not. Words that are inflected for word classes (e.g., articles behave in some respects like numerals
person, number, tense, voice, and mood form an own word class, or pronouns) or do not share any consistent set of properties (e.g.,
‘verbs’, while those being inflected for case, number, and gender are the residual word class ‘particles’). Linguistics thus, shares the same
assigned to the word class ‘noun’. Hence, word classes, known since achievements and troubles of all sciences classifying their objects of
antiquity as ‘parts of speech’ (partes orationis), are morphologically research—which categories and values to choose, which species and
defined. But the literary meaning ‘parts of speech’ indicates that subspecies to allow? and most notably, the realization that all of the
the object of morphological research is also seen as a form-part of problems in doing so can be the results of historical development,
a more complex form: sentences, speech, texts. i.e., for linguistics: language change. (NB: in most of today’s linguis-
The morphological analysis of inflected word forms also reveals tic literature, the term ‘language evolution’ denotes the evolution of
a more detailed look at word classes: the faculty of human language and not the structural development
amō videō audiō dūcō
and differentiation of grammatical structures itself. Here, the plain
amās vidēs audı̄s dūcis term ‘language change’ is prevalent, although the theory of lan-
amat videt audit dūcit guage change is as manifold, complex, and controversial in detail
amāmus vidēmus audı̄mus ducimus as the theory of evolution in biology itself. See for an introduction
amātis vidētis audı̄tis ducitis
McMahon, 1994.)
amant vident audiunt ducunt
The introduction presented morphology as the linguistic study
When comparing the paradigm of the verb amō with those of of the structure and function of word forms, and Section 1 illus-
videō ‘to watch, to look’, audiō ‘to listen, to hear’, and dūcō ‘to lead’, trated the structural component of linguistic morphology: the
it becomes obvious that the general paradigmatic pattern of verbs is study of morphemes and the systematic changes when word forms
present in all of them: a lexical morpheme is followed by a specific are altered by inflection, derivation, and compounding. Section 2
set of endings indicating person, number, tense, voice, and mood. will now examine the notion of ‘function’.
But equally, the stem ends in different sounds (ā, ē, ı̄ and consonant).
As these differences prove systematic in a wider comparison of verb
paradigms, we arrive at a morphological definition of subclasses 2. The function of word forms: values and systems
within the class of verbs: the inflection classes of the a–, e–, i– and
consonant conjugation of Latin verbs. As in biology, linguistic morphology does not make do with a
In derivation, an affix is attached to a lexical morpheme, such as mere description of the general shape of an object and its elements.
{–ness} to calm to form calmness. These combinations are subject Simultaneously, the description determines the specific relations
to specific restrictions on the part of the derivational base and the of these elements and ultimately the complexity of the object of
44 L.E. Zeige / Zoologischer Anzeiger 256 (2015) 42–53
research. The focus on relations and complexity is key to under- sentence. On the lexical axis, a lexeme is chosen from all avail-
standing the functional component of morphology in linguistics. able lexemes denoting actions, say love, hate, paint, kiss, etc. In
Thus, the notion of ‘function’ is not about the communicative func- the fabric of all verbal meanings, the value of love lies in the fact
tion of speech in interaction, or of a word meaning something or that it does not mean ‘to hate, paint, kiss’ etc. On the paradigmatic
having meaning for somebody. Those are the objects of semantics axis, the verbal lexeme comes with a full paradigm. But in order
and pragmatics. to produce a meaningful sentence, only the third person singular
Section 1 showed examples of inflected, derivated, and com- present tense indicative active is selected. Again, it is selected in
pounded word forms, like amō, calmness, and windmill. Following contrasts to all the other options. It is not chosen because it means
up with grammatical school knowledge, Section 1 also used the ‘(he) loves’ but because it does not mean everything else. The ‘func-
labels, usually put onto certain word forms, like ‘1st person singular tion’ of a word form is its difference, triangulated on the lexical,
present tense indicative active’ to amō. These terms not only serve paradigmatic, and syntagmatic axes, to any and every other ele-
as a symbol for the form, but also signify its grammatical function. ment in a language. Early structuralism introduced the term ‘value’,
Take the following example: with emphasis on the individual element (valeur; de Saussure, 1983
[1916]: 106–120), and later structuralism used the term ‘function’
Lūcius Aureliam amat.
emphasising the relations: “[. . .] une structure est par définition
The specific meaning of this Latin sentence is brought about un tissu de dépendances ou de fonctions” Hjelmslev (1942: 31).
by several means: the lexemes Lucius, Aurelia, and Amare The capitalization love-3sg.Pres.Ind.Act betokens this function. As
provide their lexical meaning, and the interpretation of who-does- a side-note it may be added how the notions of ‘function’, ‘struc-
what-to-whom is provided by the syntactic placement and the ture’, and ‘system’ are interwoven in classical structuralism. The
morphological shape of the word forms. Latin syntax allows in interdependency of all relations in a language makes ‘function’ syn-
principle for all permutations of these three words to express the onymous to the (grammatical) ‘structure’ of that language, and the
meaning of ‘Lucius loves Aurelia’: Amat Lucius Aureliam, Lucius totality of all relations makes ‘structure’ synonymous to ‘system’.
amat Aureliam, Aureliam amat Lucius, etc. The nominative singu- As in biology, form/function pairings are not unambiguous. On
lar masculine ending –us marks Lucius as the person-who-loves, the one hand, one grammatical morpheme can express several
the accusative singular feminine ending –am marks Aurelia as the grammatical functions; it is then ‘polysemous’. For example, the
person-who-is-loved and the third person singular present indica- inflectional morpheme {–s} in English indicates both the 3rd per-
tive active amat indicates that one person loves at the time of son singular indicative active of verbs (sings) and the plural of
reference without any concessions to the speaker’s personal beliefs. nouns (songs). On the other hand, at the same time there are often
Given the ‘real world’ situation where (1) somebody (2) loves (3) several strategies available of how to mark a grammatical cate-
somebody, you can imagine a syntactic pattern with three empty gory, for example, PastTense: by adding a segment (walk-ed), by
slots emerging, more or less rigorously determined in its sequence stem-modification (cling—clung), by suppletion (go—went have dif-
by language type, sentence type, and other factors. This horizon- ferent roots), or without applying any means of overt marking at all
tal axis is called the ‘syntagmatic’ axis. In the case of Lucius loving (cut—cut). Even when the morphological strategy is the same, the
Aurelia, the appropriate three lexemes Lucius, Aurelia, and Amare grammatical morpheme expressing the same grammatical func-
click into place. This can be called the ‘lexical’ axis. As each lexeme tion can take several forms; these forms are then ‘allomorphs’. A
comes with the full paradigm of word forms, the paradigms can good example is Plural marking in German: Bauer-n ‘peasants’,
slide up and down in their respective syntagmatic position—this Mensch-en ‘humans’, Kind-er ‘children’, Hund-e ‘dogs’, Kino-s ‘cine-
vertical axis is called the ‘paradigmatic’ axis. But in order for the mas’, Fahrer-ø ‘drivers’.
linguistic representation to make sense, the only meaningful posi- Form/function parings not showing one-to-one correspon-
tion of the paradigm of Lucius is to lock in place with the nominative dences and rendering language ‘illogical’ in its structure, marks
singular box, for the paradigm of Aurelia with the accusative sin- languages as dynamic objects. Their structure is determined by
gular box and for amare with the 3sg.Pres.Ind.Act box. The result function, usage, variation, and history. Form/function pairings are
is the sentence Lucius Aureliam amat (Fig. 1). subject to change. For example, the inflectional morpheme {–(e)st}
The illustrative description of ‘clicking’ and ‘sliding’ symbolizes used to mark the 2nd person (as in the 2nd person singular past
a process of selection. Let us take the example of the verb in this tense of will: thou wouldst), but this form is archaic. The mor-
pheme {–er} can be derivational (to sing—singer) or inflectional
(tall—taller), which gives rise to two interpretations: there is either
one polysemous morpheme, or there are two homonymous mor-
phemes. In fact, the historical viewpoint supports homonymy, as
different roots for the two {–er} are attested: the Old English agent
suffix {–ere}, related to and/or borrowed from Latin {–arius} in
the first case and the Old English comparative marker {–ra} in the
latter case. But there is also a strongly held position in linguistics
that denies explanatory value to historical arguments and favors
exclusively the synchronic approach to the inner complexity of any
individual language system: if it sounds alike today, it is the same.
From a theoretical point of view, the notion of ‘function’ exhibits
interesting similarities between biology and linguistic morphol-
ogy. In its classical, structuralist definition, grammatical ‘function’
is an internal function of elements in relation to the system. How-
ever, there are several newer theories of grammar which emphasize
the external functions that grammatical structures perform in
actual communication, e.g., Michael Halliday’s ‘Systemic Functional
Grammar’ (Halliday, 1990) or the works of ‘Danish Functional
Grammar’ which integrate pragmatics and discourse into the foun-
Fig. 1. Word forms and the syntagmatic, paradigmatic, and lexical axes. dations of classical structuralism (Engberg-Pedersen et al., 1996).
‘Function’ in biology basically allows for the very same dualism. The terms species, forma, and figura, surfacing in the quotation
It is my understanding that attempts to distinguish internal and from Marius’ grammar, are distinct notions of form analysis. Their
external functions terminologically have been brought forward Greek counterparts are: /eidos ‘species’, оϕ/morphe
´ ‘form’,
(‘function’ vs. ‘biological role’, see Bock and von Wahlert, 1965), and ˛/schema
˜ ‘figure’. Maybe with the exception of Marius, it
but are not generally accepted (Scholtz, 2013: 33). is rarely found that all of these terms appear together in one defi-
When discussing ‘function’, it seems also relevant to note that nition. As they all denote closely related concepts of form, some of
Richter and Wirkner (2014: 340) recently proposed to adopt the them seemed dispensable at times; they partly appear interchange-
linguistic term ‘morpheme’ to evolutionary morphology. They able or swap their meaning. Most commonly, they are distributed to
construe ‘morpheme’ as a purely descriptive term, denoting mor- certain word classes: for example, the distinction between inflec-
phological features. In linguistics, in contrast, morphemes are tional and derivational morphology, although established by Varro
defined as the smallest functional units. In dog-s, the appearance of (1st ct. BCE), was generally not made by the Late Latin gram-
the plural–s causes the plural meaning. The morpheme {–s} is the marians. Donatus, for example, uses forma for both the inflected
bearer of the grammatical function Plural, hence, a ‘plural mor- and derivated forms of verbs and figura for both the inflected
pheme’. Its specific graphic or sound quality is morphologically and derivated forms of nouns. In contrast, Priscian differentiates
irrelevant (that is to say, it is a matter of other linguistic disci- the derivated forms of all parts of speech as species derivativa (as
plines: phonetics and graphematics). A specific form of a morpheme opposed to species primitiva, ‘non-derivated forms’) and uses forma
is called a ‘morph’. In this sense, any foot-like entity of an organ- and figura for the inflectional forms of verbs and nouns, respec-
ism would belong to the same morpheme Foot, but the specific tively. Compounding, which the grammarians only identified in
shapes and forms of feet in different organisms would be individ- nouns and which is generally not a central feature of Latin, usually
ual morphs. To what extent the terminological difference between falls into the category figura.
‘morpheme’ and ‘morph’ could be useful to biology, is not up for These terms have a long history, spanning from the classical
me to discuss. period of Greek philosophy to early modern Europe. They appear
in the first grammatical treatises by classical Greek and Hellenist
philosophers mainly as part of epistemological reasoning and phi-
3. Foundations losophy of the mind. Mediated by the Stoic school, they seep into
the first autonomous grammars written by Hellenist grammarians
In the two preceding sections, inflection, derivation, and com- where they become technical terms of form analysis. Especially the
pounding were presented as contemporary concepts of linguistic aforementioned ‘Tekhne’ formed the terminological and categor-
morphology. Interested readers will have noticed similarities to ical framework for morphology in Greek and Latin grammatical
biological methodology and theory. This section will trace the studies and can be credited with putting morphology ante lit-
origins of morphological interest in Ancient Greek and Latin gram- teram at the heart of ancient grammatical theory. Later, the Roman
mars, thereby unfolding the common roots of morphology in grammarians adopted the terms to the description of Latin and
biology and linguistics. later, in East Rome and Byzantium, continued also the grammat-
Ancient grammars follow a hieratic composition—organized ical tradition of Greek. The Latin grammars, Donatus and Priscian
according to parts of speech (usually eight), each word class is in particular, are then in use throughout the Middle Ages. It is overly
presented with a definition, their inflectional and derivational simplified to claim that the Greek and Latin grammatical traditions
categories, partly discussing problems of classification, usually of Byzantium and the Middle Ages were little innovative and lack-
exemplified by Greek or Latin word forms, respectively. These ing originality. In fact, grammatical writing developed in a closely
works span anything from a few pages to multi-volume books. knit network of classical texts being copied, compiled, and com-
Judging from their impact on medieval and Renaissance grammat- mented on as well as inspiring autochthonous works (Robins, 2000:
ical writing, the most influential ancient grammars were the Greek 58ff.). These Latin grammatical treatises and grammars are in use
‘Tekhne grammatike’, generally ascribed to Dionysios Thrax (2nd ct. up to and throughout the early modern period. Scholars who were
BCE) and reaching its canonical form by the fourth century, Aelius trained in Latin grammars then used their knowledge to compose
Donatus’ ‘Ars minor’ and ‘Ars maior’ (4th ct.), which dominated the grammars of the modern European languages. In the beginning,
early and High Middle Ages, and Priscian’s ‘Institutiones grammat- these grammars were also written in Latin, but later used the lan-
ica’ (6th ct.). In contrast, the 3rd century grammar by the otherwise guages of the people. When doing so, they either kept the original
unattested Marius Plotius Sacerdos (Keil, 1874: 415–546) was nei- Latin terms, or introduced vernacular translations, or both. For
ther particularly ingenious nor influential, but it is interesting in example, in Christian Gueintz’ (1592–1650) ‘Deutscher Sprachlehre
another respect. It exhibits a large set of descriptive terms concern- Entwurf’ (1641) the difference between species and forma/figura
ing the morphological structure of word classes, otherwise only appears as the difference between Art and Gestalt (Gueintz, 1978
found across the definitions of several word classes or across dif- [1641]: 24). In general, it is safe to say that there is a direct line
ferent grammarians. Some of these terms will sound familiar to of tradition from ancient philosophy and the first grammars of
comparative botanists and zoologists—and this for good reasons. Greek and Latin to the modern linguistic occupation. It committed
When we take the verb as an example, Marius lists the following grammatical study to a form-oriented, morphological approach to
terms: language.
verbo accidunt VIIII, forma, qualitas, genus, quod dicitur adfec- It is not by chance that species, forma, and figura became rel-
tus vel species vel significatio, figura, numerus, modus, tempus, evant terms in both biology and linguistics. They are rooted in a
persona, coniugatio (Keil, 1874: 429). philosophical tradition ranging from Plato to Aristotle and the Hel-
lenist Stoics. At first, Plato (428/427–348/347 BCE) divided matter
By his reckoning, the verb has nine ‘accidents’ (accidentia), i.e., ( /hyle) and form ( /eidos; оϕ/morphe),
´ proposing that it
morphological properties. Number, mood, tense, person, and the is the form imposed on the matter which gives matter its identity.
inflection classes of the verb are in their core compatible to today’s Regarding the form, he proposed the concept of ideal forms present
terms. ‘Quality’ captures the distinction of finite and infinite forms in every object of the same category. The essence ( /to
of the verb, while genus refers to voice and is therefore, in the case ti ên einai ‘the what it was to be’) is permanent, unalterable, and
of the verb, synonymous to adfectus and significatio. The remaining eternal, whereas, the objects are non-permanent and variable. But
three terms (forma, species, figura) will be the focus of this section. matter and form cannot be detached from another. The essence of
the thing can therefore be indicated by its form. Consequently, the yet sufficiently separated from a genealogical perspective. But the
study of objects aims at the ideal form, but is based on observation ambiguity caused paradoxes which in a continued mutual discus-
and the formulation of laws about the external form (which in the sion of empirical evidence and theoretical foundations established
case of objects of the real world is actual physical form). Or, as Plato an autonomous theory of genealogical transformation. One natu-
put it into the mouth of Timaeus: rally expects pre-Darwinian taxonomical work to show the same
First then, in my judgment, we must make a distinction and ask, ambiguity, causing the same disciplinary development. (To what
What is that which always is and has no becoming; and what extent ‘essentialist’ practice is carried over into 20th century biol-
is that which is always becoming and never is? That which is ogy, is another matter.)
apprehended by intelligence and reason is always in the same Regarding the first insight, the holistic perception of objects, the
state; but that which is conceived by opinion with the help of idea of essential and accidental properties of form brought about
sensation and without reason, is always in a process of becoming the practice of defining word classes of the type “verbo accidunt:
and perishing and never really is (Timaeus 27–28). . . .” as well as the holistic approach to grammatical structure that
has been called “word and paradigm” (Hockett, 1966 [1954]): as
In Aristotelian philosophy, this was reinterpreted, expanded, can be inferred from the absence of terms for ‘root’, ‘stem’ or ‘affix’
and specified, especially the difference between essential and non- (let alone ‘morpheme’), the ancient grammarians were lacking the
essential properties of things. Aristotle (384–322 BCE) claimed “building-block mentality” (Law, 2000: 78f.) that is common to
that in everything, there are elements existing only because large parts of contemporary morphology. In contrast to today’s
they are sustained by other elements, while other elements exist atomistic approach presented in Section 1, focus was put on indivis-
autonomously. He came to distinguish in each thing the essential ible word forms in their systematic relations within paradigms. The
element which has being in itself (ens in se) from the accidental ele- forms gathered in a paradigm represented the essential nature of
ments which have being because they exist in another thing (ens in the item. It is also important to note that, from the beginning, species
alio) and modify it. In detail, he laid out the ten properties of things and forma/figura did not serve as descriptive terms for the perceiv-
(Categories 1b25–2a4; Metaphysics IV, 1089b23–1090a4), of which able physical elements of language (the form of sound waves, or the
the first was the essential property ‘substance’ and the other nine form of written texts), but for the abstracted functional structure of
where the accidental properties of form—‘quantity’, ‘quality’, ‘rela- grammatical categories. The holistic objects of linguistic study were
tion’, ‘place’, ‘time’, ‘position’, ‘state’, ‘action’, and ‘affection’ (cf. the word forms as part of their paradigms. This made the word the
some properties in Marius’ definition of the verb!). ‘Quality’ is of linchpin of linguistic reflection, based on a study of form and shape,
particular interest here, as one of the ‘sorts of quality’ is form: from antiquity and for about 2000 years onwards.
The fourth sort of quality is figure [schema] and the shape [mor- As we have seen, the morphological tradition in linguistics
phe] that belongs to a thing; and besides this, straightness and emanates from the same philosophical source as in biology. The
curvedness and any other qualities of this type; each of these terms species, forma, and figura are pivotal to the structural and
defines a thing as being such and such. Because it is triangular or form-oriented approach and anticipate the much later and artificial
quadrangular a thing is said to have a specific character, or again term ‘morphology’, in that they are epistemological notions relat-
because it is straight or curved; in fact a thing’s shape in every ing individual phenotypic properties to abstract categories (see for
case gives rise to a qualification of it (Categories 8b25–11a39). a discussion in biology, Töpfer, 2013). Goethe’s original intention,
signified by the creation of the term ‘morphology’, was to pro-
Hence, we have ‘form’ meaning the general shape of an object duce a hierarchy of the sciences and to pinpoint a fixed position
as opposed to its matter, and ‘form’ as a rather graphic property for a universal science of the properties of form: a science of the
and a descriptive term. Being rather concrete in the beginning, the “Gestalt sowohl in ihren Teilen als im ganzen, ihren Übereinstim-
difference of essential and accidental properties carried over into mungen und Abweichungen ohne alle anderen Rücksichten” (von
all facets of the world, from physical objects to abstract concepts. Goethe, 1962 [without year]: 123; see also Wildgen, 1984). This
But the idea was left untouched that the common characteristic envisioned universal science was never achieved. Given their com-
of all things belonging to a particular category, defining them as mon theoretical ground and the central position of morphological
members of that category, could only be found by the observation studies in their respective disciplines, it seems that morphology is
and description of the external properties, most notably the objects’ rooted most deeply in biology and linguistics. They could hence be
form, figure or shape. Thus, Aristotle’s essentialism informed two assigned the task to propagate and develop morphological thinking
important insights. First, it gave rise to his doctrine that things are as part of the structural sciences.
not mere aggregates of independently existing parts. The whole has
priority over its parts and is not adequately defined as the sum of 4. Contributions of morphology to the language sciences
its parts. Second, it allowed for a non-destructive view on varia-
tion and change: accidental form is a dependent property. Forms Apart from making structural properties of language accessible,
modify a subject or substance, but do not alter its identity. In the morphology has also contributed greatly to the understanding of
course of time, the same thing can change its form substantially how languages are related, by adding morphological classification
and still be the same thing. These two insights are common to the to language comparison.
understanding of form in both biology and linguistics. Language comparison in the early modern period showed an
Regarding the second insight, the considerable debate in biol- interesting dualism that is amongst other things expressed by the
ogy on pre-Darwinian ‘essentialism’ (Winsor, 2003; Amundson, name under which the study of species and forma/figura traded in
2005; Hull, 2006; Müller-Wille, 2007) is surprising. Doubtlessly, it most early modern grammars: ‘Etymologia’ (e.g., Schottel, 1967
became the rationale of taxonomy to observe and describe the form [1663]: 224). It signified primarily the descriptive study of words
features of specimens and to categorize the specimens according as it referred to the changes a word would suffer from processes of
to abstracted form principles. This is equally true for linguistics inflection, derivation, and compounding. In that sense, an inflected
at the time. But—as will be discussed in more detail in Section or derived word was ‘related’ to its root. Today of course, the
4—linguistic study before the 19th century examined and com- term ‘etymology’ both signifies the scientific study of the history
pared linguistic specimens with a deeply felt intuition that the of words as well as the history of a word itself, its origins, and
similarities and differences that facilitated classification also indi- how its form and meaning have changed over time. Today’s mean-
cated kinship. Certainly, the synchronic classificatory work was not ing was present in the earlier one, too. The ‘creation’ of a lexeme,
the question of how a word came to name an object or action, the (2) Latin habēre – German haben
manipulation of its form (with subsequent shift in meaning) and (3) English cheap – German kaufen
the developmental dimension of words as entities of an evolving (4) English tariff – Arabic ta rı̄f
language conflated in pre-historicized language description. The
separation into morphology on the one hand, and etymology in the The words for ‘father’ in (1) are taken from languages spanning
modern sense of the word on the other hand, started not before from Northern Europe to the South and to the East as far as India,
the end of the 18th century, as can be seen in Johann Christian also backwards in time. Today, being established by almost 200
Adelung’s (1732–1806) ‘Umständliches Lehrgebäude’ (1782), the years of scientific research and a large scale comparison of word
last of the great learned grammars of German before the outset of forms of several languages, there is a set of laws of sound differen-
the modern language science: a lengthy discussion of the recent tiation available, allowing to propose the ‘Indo-European language
state of etymology (Adelung, 1971 [1782]: 177–244) leads from family’, of which the word forms given in (1) sample an illuminative
the origins of roots in the imitation of sound to the morphological probe. The unconditional sound law, for example, of /p/ changing
processes of inflection, derivation. and compounding, and finally to into /f/, is one of the set of laws defining the partition of the Ger-
differentiation into dialects and languages. manic language family from the rest of the Indo-European family.
The structural study of languages and dialects in the early mod- If the comparison of sounds is based on roots of common etymol-
ern period was sufficiently detailed to enable initial classificatory ogy (‘cognates’), the idea of phonological comparison is not wrong
work. However, these works exhibited the very same dualism and has proven fruitful to an extent that cannot be illustrated here
in that they were as much classificatory systems as they were in further detail. To this day, it is the methodological basis of our
hypotheses on descent. Earlier and to a greater extent than in biol- understanding of language differentiation, kinship, and descent,
ogy, structural similarities between languages were interpreted not only within the Indo-European language family.
as indications of descent and subsequently fostered comparative However, the example in (2), German haben ‘to have’ and Latin
work since at least the 16th century. But it needed the theoreti- habēre ‘to hold, carry, include, have’, is a different case. They,
cal and methodological developments put forth by biology in the being seemingly similar in their form and having a clear over-
19th century to turn from etymological speculation to systematic lap in meaning, are likely to be shuffled together. But the laws
analysis. of sound differentiation indicate that for example German /h/ and
For example, Fig. 2 is to be found in Justus Georg Schottel’s ‘Aus- Latin /h/ cannot be etymologically identical sounds. When applying
führliche Arbeit von der Teutschen HaubtSprache’ from 1663. It the full set of relevant sound laws to these words, German haben is a
is a graphic account in the style of a ‘système figuré’, classifying descendent of the proto-Indo-European root *kap- ‘to grasp’ (Kluge,
languages and dialects of Central, Northern, and North-Western 2002: 380; Pfeifer, 1997: 490) and hence related to Latin capere ‘to
Europe. In today’s terms, the brackets subsume the languages in seize, to grasp’, whereas, Latin habēre is a descendent of PIE *gh abh -
the right column into three groups—North- and Western-Germanic ‘to give, to receive’ (Kluge, 2002: 335; Pfeifer, 1997: 405f.) and thus,
languages (top) and the difference between Low German (mid- related to German geben ‘to give’. In other words, both the form and
dle) and High German (down) dialects. It is a rough account of the meaning of words can be deceiving.
the Germanic language family. With today’s classificatory knowl- In (3), both the English adjective cheap ‘low priced, poor in qual-
edge, of course, some classifications are inaccurate. For example, ity, of little worth’ and the German verb kaufen ‘to buy’ descend
the inclusion of Scottish, Irish, and Welsh as well as Gothic into from a common root: early in the history of the Germanic lan-
the uppermost bracket, the first three belonging to the Celtic languages, the Latin noun caupo ‘innkeeper, merchant’ gave rise to a
guage family and the latter is the only documented east-Germanic Germanic verbal base *kaup-ō- ‘to buy, to trade’ (Kluge, 2002: 479;
language. But given the age and the methodological and theoret- Pfeifer, 1997; 639f.), from which a number of adjectives and nouns
ical level of development, Schottel’s account seems remarkably derived: the Old English noun cēap (‘cattle, purchase, sale, busi-
well informed. But most interestingly, the passages surrounding ness’) and verb ċēapian (‘to bargain, to trade, to contract, to buy, to
the figure both comment on a synchronic classification that Schot- bribe’) gave rise to the adjective cheap and later were lost; in Ger-
tel portrays as an arrangement of ‘dialects’ with different features man the verb kaufen still exists as well as the derivated nouns and
(zertheilet ‘divided’, §9, p.151; §12, p.152) and the continuous form- adjectives (e.g., der Kauf ‘purchase, acquisition’, der Käufer ‘buyer’,
ing, separation, and coming into existence of these ‘dialect’ (“immer käuflich ‘purchasable’). In subsequent semantic change, the English
mehr und mehr zertheilet, ausgebreitet und neue Mundarten darin root narrowed its meaning to a specific part of the original meaning
entstanden und aufgebracht worden” §9, p.151). (‘to buy’ > ‘to bargain’ > ‘inexpensive’, and partly > ‘of little value’).
Classifications of that kind are usually based on a comparison of Systematic changes in meaning and shape make inheritance some-
word forms. If two languages exhibited a certain amount of words times hard to detect.
with similar meaning and form, they would be regarded as related. The Arabic noun in (4) derives from a verb meaning ‘to notify’,
The methodological and theoretical level, on which the compari- hence meaning ‘notification’ (Kluge, 2002: 906). The noun has been
son of languages was executed, is originally informed by personal a loanword in European languages, since the Italian Renaissance
language experience and a subjective perception of differences, merchants of Venice picked it up in their trading posts in Ara-
peppered with anecdotal evidence. The perception of form is mostly bia. It has since then modified its meaning through continued use:
impressionistic. The most palpable variables of difference were the German Tarif and French tarif ‘rate, scale, fee, price’; English tar-
sound qualities, which at the time were not separated from the iff additionally ‘financial duty, tax, fee, toll’. The example shows a
visual impression of letters in written and printed texts. Often, case of lexical borrowing. It is matter of debate, to what extent,
vague analogies in the letter/sound structure of presumably com- on which paths and under which circumstances language contact
parable words gave rise to all sorts of speculation on kinship and leads to language change. There is evidence that language contact
descent, something that later would be deprecatingly called Ety- is not restricted to loanwords; but phonological, morphological,
mologisieren ‘etymological sleight of hand’. Consider the following syntactical, textual, and pragmatic features can also diffuse from
examples: one language to another. But morphological features, it seems, only
transfer when nested in lexical borrowing (Winford, 2013: 175). In
phenomena of language contact, linguistic structures behave rather
(1) Vedic Sanskrit pitár – Greek ˛
´ – Latin pater – English father culturally, exhibiting characteristics uncommon to biology. Medi-
– Swedish fader – German Vater ated by bilingual language users, two or more species of languages
Fig. 2. Early language classification (Schottel, 1967 [1663]: 153).
coexisting in the same area can exchange features and over time illustrate the difficulties of classificatory work that seeks to
become more similar, regardless of whether they are genetically determine essential categories by the analysis of accidental form,
related as members of the same language family, or not. Usually without a genealogical theory at hand (see Scholtz, 2010: 47f. for
in these constellations, one species is a dominant giver-language. biology). The discontent led to two developments in language com-
In extreme cases, the process may replace one species completely parisons. First, a methodological sharpening of sound comparisons,
(‘language shift’ of a population towards one language resulting in culminating in a school of linguists called the ‘neo-grammarian’
the ‘language death’ of another). In particular contact situations, (1876–1900), which proclaimed their rigorously mathematized
the languages involved can also merge their features into a linguis- positivist sound laws as the only reliable foundation of language
tic structure which re-stabilizes as a new grammar clearly different genealogy. Second, is the inclusion of morphological properties
from the base languages (e.g., Creole languages). Another biological into the comparison. While the neo-grammarians represented first
impossibility is the introgression of fossil characters into a living and foremost a methodological advancement, it was the latter that
organism, which is in linguistics found in the creation of learned advanced a linguistic theory of genealogy.
terminology from Ancient Latin or Greek roots. Phenomena of lan- Morphological comparisons, especially the criterion of present
guage contact present not only a problem to language classification or missing inflectional morphology, have been identified in
in that they dilute the computation of genealogies. Ultimately, scattered works as early as the 17th century (Meier, 1996: 465ff.).
they challenge the appropriateness of the idea of linguistic descent The following account will be restricted to the works of two individ-
(Krämer, 2014: 94). uals, particularly significant to the role of morphology—Friedrich
Problems of the kind demonstrated by the examples in Schlegel (1772–1829) and August Schleicher (1821–
(1)–(4) accumulated throughout the 17th and 18th century. They 1868).
Schlegel’s ‘Über die Sprache und Weisheit der Indier’ (1808) is bly most iconic illustration of a linguistic family tree (1861), and he
a programmatic writing and a most influential work. Consider the demanded the application of Darwinian ideas to language evolution
following brief passage: (1863). The basis of all three innovations was the deep belief in lan-
guages as natural organisms. Whereas, grammarians in the Age of
Jener entscheidende Punkt aber, der hier alles aufhellen wird, ist
Enlightenment presumed ‘grammar’ to be mechanistically and log-
die innre Structur der Sprachen oder die vergleichende Gram-
ically decomposable in the way of a learned ‘arti-ficial’ profession
matik, welche uns ganz neue Aufschlüsse über die Genealogie
( έ ˛˛ /ars ´ grammatica), the vast complexity and
der Sprachen auf ähnliche Weise geben wird, wie die ver-
flexibility of grammatical structures now nourished a perception
gleichende Anatomie über die höhere Naturgeschichte Licht
of organic relations and unconsciously and ingeniously function-
verbreitet hat (Schlegel, 1995 [1808]: 28).
ing inner workings of language. In Schlegel’s orientation towards
In the sense discussed above, it demands a shift of attention comparative anatomy, the idea of ‘language as a natural organism’
from roots to morphology—towards the “most inner structure was allegorical and never actually expressed. But with Schleicher, it
and grammar”—to reveal the true homologies and thus, to prove coalesced with the generally positivistic disposition of modern sci-
kinship. Schlegel reckons that morphology is much less suscepti- ences towards the end of the 19th century and advanced the belief
ble to the disturbing influence of language contact (Einmischung) in the lawfulness of linguistic structure and change (cf. Schleicher,
than a language’s vocabulary. The analytical chapters on mor- 1863: 6–12). The analogy informed the method and ultimately
phology are mainly based on declension and conjugation and became a true ontology (Cassirer, 1945). Even today, the popular
demonstrate—although obsolete in many details today—the virtues language use of ‘language decay’ or ‘languages dying’ reflects the
of structural comparisons as opposed to cognate-based compar- 19th century approach to languages as natural organisms.
isons (Chapter 2 and 3, pp.28–43). The other noteworthy aspect We said above that morphology does not only describe elements
is how explicitly the picture of language genealogy is informed of linguistic form, but always seeks to define the relationships
by comparative anatomy. Schlegel is documented to have been between these elements and hence, the complexity of a language.
in Paris around the year 1803, where he witnessed the impact of Schleicher’s introduction of the term Morphologie—although it had
palaeontologist George Cuvier’s (1769–1832) ‘Leçons d’anatomie been used before by Max Müller in 1854 (Morpurgo-Davies, 1998:
comparée’ (1800–1805). He adopted the approach to linguistics: 221 fn19)—presupposes much of this objective:
a comparison produces similarities which in turn lead to the
Für die lere von der wortform wäle ich das wort ‘morphologie’,
postulation of kinship and descent. Thus, Schlegel binds the ques-
nach dem vorgange der naturwißenschaften, weil ‘formlere‘
tion of language kinship to the study of grammatical structure.
für specielle morphologie verbunden mit functionslere der
The orientation towards comparative anatomy also entailed that
beziehungslaute bereits im gebrauche ist [. . .] (Schleicher,
comparative linguistic study should commit to a non-speculative,
1859: 35).
descriptive, and empirical approach and adhere to the accuracy
exhibited by comparative anatomy. As Hutton (1995, xii) points He established a difference between (a) a universal ‘mor-
out, “the actual linguistic analysis is concerned only with the estab- phology’ in terms of the comparative study of word forms and
lishment of formal relationships (. . .). ” In the course of time, this exerted with the rigour of the sciences—meaning first and fore-
did promote the profile of linguistics as an autonomous academic most biology—and (b) a ‘special morphology’ in terms of the study
discipline. But Schlegel’s programmatic aim was to prove the ori- of word forms of a particular language, obviously thought of as tra-
gin of all Western language, literature, and philosophy in Indian ditional grammar-writing. Thus, Schleicher modernised the notion
(Vedic) antiquity (cf. Schlegel, 1808: 173–195). In that, his approach of ‘form’. In (b), ‘form’ is continued in its ancient and mediaeval
exceeded a mere classification of languages and was historical in tradition—forms are ‘forms of something’, modifications of some-
nature (Trabant, in press). One is also bound to point out that thing essential. In (a), by contrast, ‘form’ is an abstracted, universal
Schlegel sets the Indian origin in particular against other East- form type, which characterises a certain language as a member of
ern and Western languages, their speakers and cultures (Messling, a class. These ‘forms’ are structural categories in their own right
2013: 36–39). within the fabric of grammar which makes Schleicher’s morphol-
To August Schleicher, modern linguistics owes three inno- ogy part of Goethe’s universal study of form.
vations. He introduced the term ‘morphology’ for the study of On the basis of the universal morphological properties of the
structure and function of word forms (1859), he drafted the proba- word, Schleicher attempted to produce a new model of linguistic
Fig. 3. Family tree of the Indo-European language family (Schleicher, 1861: 7).
kinship by establishing a new natural science of language which he

called ‘Glottics’:
Dise selbst ist ein teil der naturgeschichte des menschen.
Ire methode ist im wesentlichen die der naturwißenschaften
überhaupt; [. . .] Eine der hauptaufgaben der glottik ist die
ermittelung und beschreibung der sprachlichen sippen oder
sprachstämme [. . .] (Schleicher, 1861: 1).
To ascertain the ‘kin’ and ‘roots’ of languages, Schleicher’s
approach exceeds the purely phonological approaches sketched Fig. 4. X-bar-scheme of syntactic structure.
above that infers descent from the differences in sounds. He relied
on morphological comparisons which collect and classify linguis-
tic forms in the sense of comparative biology. The results led the word and towards higher aggregated structures, and second,
Schleicher to draw graphic models of the Indo-European language ‘big data’.
family. These visualizations inherited the dualism of classificatory In linguistics today, it is viewed as the major capacity of
method and genealogical interest. In Schleicher’s earlier family tree the human language faculty to transform complex simultane-
(Schleicher, 1853), the notion of descent was only loosely sepa- ous thoughts into linear acoustic or graphic expressions. Take for
rated from a comparative classification. After reading Darwin’s ‘On example, a situation of somebody drinking coffee. An observer or
the origin of species’, Schleicher effectively tied historical linguistic participant possesses a mental representation of the situation, a
thinking to the process of evolution (Schleicher, 1863). Thereby, the concept. It includes three entities at the same time: (1) an action,
family tree model turned into a graphic account of linguistic his- the ‘drinking’, (2) an identifiable individual as an agent of this
tory and descent charged with a new explanatory force. Schleicher’s action, a ‘drinker’, say George, and (3) the object subjected to the
family tree from his main work, his ‘Compendium’—published action, a ‘drinkee’, the coffee. To arrive at a pronounceable utter-
around the same time in 1861 (Fig. 3) and in several reprints ance, these entities have to be transformed into a linear syntactic
onwards—became the iconic illustration of this converted view on base structure that holds generally attributed slots for actions, its
language (Zeige, 2013). agents and objects (Fig. 4 presents a traditional account).
Schleicher was surely the most rigorous proponent of biolo- Out of the base structure, all possible conventionalised struc-
gism in linguistics and his ideas were both challenged from the tures of sentences in language usage can be produced according to
outset and in parts revoked by his immediate successors. In partic- further conditions: e.g., George drinks coffee to express the habitual
ular, the school of the ‘neo-grammarians’ stood rather indifferent process, George drinks the coffee to emphasise the selection, or The
to the notion of language organisms, as they redirected their labour coffee is drunk by George to tone down the agentivity of the drinker,
towards single morphological forms as parts of paradigms. For and so on.
example, in their ‘Morphologische Untersuchungen auf dem Gebi- Putting sentence structures in the centre of attention, it is of
ete der indogermanischen Sprachen’ (six volumes, 1878–1910), lesser interest that ‘drinking’ is lexicalised as drink–, what part
Hermann Osthoff (1847–1909), and Karl Brugmann (1849–1919) of speech this word belongs to or how it is inflected. In different
presented in large detail the phonological form and sound laws theoretical frameworks, morphology is either degraded or even
of inflectional and derivational morphemes and of stem changes allocated to different grammatical layers. Early American struc-
and did not participate in the morphological interest in form types. turalism focussed on classes of morpheme sequences—with single
With its indefatigable attention to detail, their work exhibits an morphemes being “merely the special case where the sequence
unbreakable positivist belief in the explanatory power of empirical consists of one” (Harris, 1946: 165)—levelling the distinction of
data and deduced laws. It is characteristic for the Indo-European whether such a morpheme sequence is in fact a word or a syntac-
studies (Indogermanistik) towards the end of the 19th century. tic construction. Within the framework of Generative Grammar, the
In the period covered by the works of Schlegel, Schleicher, and properties of words are explained by recourse to general grammat-
the neo-grammarians takes place the institutionalisation of linguis- ical principles which apply to syntax and morphology equally (e.g.,
tics as an autonomous academic (as opposed to learned) discipline. in Selkirk, 1982). In the influential framework of ‘Distributed Mor-
Chairs in comparative linguistics (“Allgemeine Sprachkunde”, “All- phology’ (Halle and Marantz, 1994), complex words and complex
gemeine Sprachwissenschaft”, “Grammaire comparée”, “Anthro- phrases are produced by a single so-called ‘engine’. Also the role
pologie linguistique”, “Comparative Philology”; cf. Ziegler, 2000: of lexical elements is drastically reduced as the lexicon only stores
93) were established from 1821 onwards in several European bundles of semantic and syntactic features to which phonological
countries. It can be said that the methodical orientation towards content (form) is associated at a much later stage. The radical con-
biological disciplines and the inclusion of morphology into lan- cept of ‘split morphology’ (e.g., Anderson, 1982, 1992) claims that
guage comparisons enhanced the precision of linguistic study and derivation is generally handled by the lexicon, while inflection is
thus, the reputation of linguistics. It should also be noted that lin- handled by syntax. Indeed, the syntactic structures in the example,
guistics, despite its focus and diversity today, started as a historical He drinks a tea/a shake/a milk shake is the same, regardless of the
and comparative science. In being partial to biological methods and object noun being a lexical root, a derivated noun (from the verb to
theories, linguistics also followed the grand fascination of the 19th shake) or a compound. Syntax seems to be blind towards morphol-
century: ‘time’ as a category of knowledge. It proceeded from the ogy. By contrast, inflection is obviously subject to the requirements
collection of specimens to the awareness of historical processes of of syntactic rules when one grammatical category is expressed by
variation and change. several words of an expression at the same time. In Amat Lucius
puellam pulchram et ingeniosam (‘Lucius loves a beautiful and witty
girl’), the object case is marked three times. The accusative sin-
5. Modern challenges gular suffixes are inflectional morphemes and hence entities of
morphology, but the function is at the same time syntactical, as the
As far as can be seen from the inside of an evolving discipline, lin- agreement marks the unity of the object’s noun phrase in its func-
guistic morphology as a self-contained subfield of linguistics faces tion as the object of the sentence. Therefore, it is rather of interest
two main challenges—first, a shifting focus in linguistics, away from which order of constituents conveys which meaning. Grammatical
structures are hence understood from syntactic structures ‘down- complex ending suggests that the dental suffix might have orig-
wards’, putting sentences and phrases into the center of linguistic inated from an autonomous verbal root: PIE *dh ē-. As a full verb
attention rather than words. Institutionally, this can be observed as (Protogermanic ), it originally meant ‘to do’ (compare English
a shift from morphological research to syntactic research, including do in He does his homework). But it was also in use as an auxiliary,
the subsequent allocation of resources. following a full verb. In this state, the construction is comparable to
The notion of the ‘word’ generally seems biased by Indo- English constructions with do marking the past tense (He did sing
European languages, that is to say, its usefulness depends on at the opera; auxiliary do + full verb sing). When the transmission of
the morphological type of a language. Take the following exam- Gothic in manuscripts began, the verbal complex had already been
ples from non-Indo-European languages: (1) in the Finnish word amalgamated into a single word form. The former auxiliary in its
taloissani, the lexical morpheme {talo–} ‘house’ is combined with inflected forms serves as an ending, here salbon + dōn.3Pl.Past: sal-
grammatical morphemes denoting the plural {–i–}, the locative bodedun. The development of the dental suffix fits into a pattern
{–ssa–} and the possessive {–ni–}, giving rise to the inflected mean- known as the ‘cline of grammaticalization’ (Hopper and Traugott,
ing ‘in our houses’. (2) Koyukon is a Native American language of 2003: 7):
Alaska. The expression means ‘The wind is
The cline of grammaticalization
moving it around’ and is analyzed as follows (Asher and Simpson,
1994: 2602): content item → grammatical word → clitic → inflectional affix
kk’o- y- ee- ’oyh A full verb (or any lexical, or content item) gradually loses
Around Wind obj imperfective To move something solid its lexical semantics and becomes a ‘light’ verb (cf. in I took a
shower, the shower is not actually ripped off the wall and carried
It contains elements that can be classified as a lexical noun
away; the expression as a whole means ‘to shower’). It becomes
(‘wind’), as a verb (‘to move something solid’), and as various gram-
a word form that expresses only grammatical but no longer any
matical morphemes. But is this expression a word or a sentence?
lexical meaning: a grammatical word, e.g., an auxiliary. Auxil-
Its translation creates an English sentence, but in phonological
iaries are rarely stressed and hence exhibit a tendency to ‘lean’
respects and with regard to inflection, the structure behaves rather
against and ultimately become bound to other elements of a sen-
like a word form. Maybe this is a wrong question after all. On the
tence (they become ‘clitics’). When these combinations present full
opposite end of the scale, there are languages seemingly consist-
paradigms, the original syntactic phenomenon of combining word
ing of nothing but uninflected words. Their syntactic constructions
forms turned into something morphological: first word formation,
are formed by the non-interchangeable arrangement of immutable
then inflection. As inflectional endings have a tendency to reduce
lexical words, as in the Mandarin Chinese example:
in phonological substance, the former complex structures gradu-
tā zài túshūguăn kàn bào ally erode, losing even their state as a syllable. For example, the
He At Library Read Newspaper remnant of the original full verb in German weak past participles
‘He is at the library reading the newspaper’ (Li and Thompson, 1981: 157)
is a simple /t/, as in gesag-t ‘said’.
In the absence of inflectional and derivational morphology and Research has, in other words, brought up several factors
paradigms, these words can only be classified according to their which diminish the prominent role of the word and its structural
syntactic behaviour or their semantics, e.g., a word substituting features—(1) the general disciplinary shift towards syntactic struc-
Library in the example above must be a noun, or words denoting tures, (2) the growing knowledge of language structures outside of
‘things’ are nouns, and words denoting ‘actions’ are verbs. But this the Indo-European language family, and (3) synchronic problems
is far from clear cut, also in languages with a richer morphology, of classification at the fringes of morphology and syntax arising
such as English: how ‘nouny’ is an abstract noun, derivated from from transitions between elements of syntax and elements of mor-
a verb (e.g., (the) love)? It certainly behaves like a noun syntacti- phology. The distinctions between morphemes, word forms, and
cally, but semantically it denotes a variety of emotional qualities phrasal structures start to blur in that morphemes and word forms
(adjectives), processes (verb), and attitudes (noun). Is a participle in some cases cannot be distinguished and some syntactic rules also
that is used as an adjective still ‘verby’ (e.g., the married man)? It govern the inner structure of word forms. Therefore, the lineariza-
certainly belongs to the paradigm of a verb, but otherwise exhibits tion of morphemes in the morphological processes of inflection,
the behavior of the nominal domain. The discussion of the ‘verbi- derivation, and compounding seems less different from the lin-
ness’ and ‘nouniness’ of roots has led to scalar approaches to word earization of phrasal elements in syntax. In fact, both can be seen
classes (see for an overview Sasse, 2001). as epiphenomenal to the necessity to linearize.
There are obviously different strategies to mark syntactic rela- In general, it became apparent that the idea of morphology in
tions and derivational meaning, and almost all known languages terms of an autonomous grammatical module was biased by the
exhibit properties of all structural types. The languages of the world structures of European languages. It is not sharply set off from syn-
differ greatly in the extent to which they make use of morpholog- tax, as it is generally not sharply set off from any other grammatical
ical means, some can be said to possess almost no grammatical module. The disciplinary specialisation into phonology (structure
morphology at all. Also, genetically related languages can be very of sounds), morphology (structure of words), syntax (structure of
different in this respect: judged from their common heritage in the phrases), semantics (structure of meanings), and lexicology (struc-
Indo-European family, English makes much less use of grammatical ture of the word-stock) has been challenged since interest grew in
morphology than, for instance, Icelandic. phenomena at their interfaces. Consequently, unifying terms have
Within one language, it is possible not only to identify cases of been introduced: ‘morphosyntax’ (the expression of syntactic func-
different structural types but also to observe single elements strid- tions by morphological means), ‘word phonology’ (word-internal
ing between syntax and morphology. To take a traditional example properties of single sounds and prosodic properties especially
(Kern and Zutt, 1977; Ramat, 1981), the preterit and preterit dependent from the morphological structure of word forms),
participle endings of the weak conjugation in Germanic languages ‘morphophonology’ (sound changes occurring within morphemes
regularly include a /d/ or /t/. They are hence called ‘dental suffixes’: when combining), or ‘word semantics’ (morphological form in
English walk—walked—walked, German hören—hörte—gehört, Nor- the context not primarily of grammatical function but meaning
wegian ringe—ringte—ringt. A particularly enlightening example is in the terms of semantics). To only pick up ‘morphosyntax’ here,
Gothic salbodedun, the 3rd person plural preterit of ‘to anoint’. Its on a descriptive level, it stands for the unified study of form and
functions of constitutive elements in utterances. Although giving characters were priced in a simple model (replace, delete, and
up on the name ‘morphology’, the study of the form and function insert = 1; match = 0) and a feature-weighted model according
of grammatical structures actually increases. It has to be noted, to phonological values (e.g., the loss of voice (/b/—/p/) was less
however, that, following the paradigm that accompanied the suc- costly than the loss or insertion of an element). The matrices were
cess of syntactic studies, much of the research on morphosyntax then treated with phylogenetic algorithms: hierarchical clustering
is executed with the aim to learn something about the human and one-dimensional MDS-plots produced dendrograms, the
word-forming capacity (Toman, 1998: 308) rather than learning additive tree model produced trees. Generally, the tested methods
something about grammatical structures. proved capable to generate meaningful chronological and dialectal
The second challenge is ‘big data’. In large scale data-bases of interpretations. But the distances differ substantially and are a
languages, annotating linguistic data with morphological informa- matter of debate, though it has to be said that historical linguistics
tion is costly. It consumes manpower, time, and money. On the usually perceives periods and dialectal differences according to
other hand, the internet, text messaging and other digital tech- isolated characteristic features. Distance measures may provide
nologies provide natural, usage-based linguistic data ready to be future arguments with which to justify or modify established
tapped (legal issues aside). The question is to what extent the classifications. The interesting point raised by Hochmuth’s account
computability of unanalysed strings of characters can replace the is yet another: a character-based comparison of phonetic or ortho-
linguistically informed analysis of morphological structures. Inter- graphic form implicitly includes some morphological and syntactic
estingly, key impulses have again been taken from biology—more information. When, for example, the three inflection classes of Old
precisely, bioinformatics—and one major focus of work is again the High German weak verbs (–ēn, –ōn, –jan) collapse into one single
language classification with subsequent genealogical interpreta- paradigm, the substitution matrix would reflect the morphological
tion. To that end, computational phylogenetic models, developed change in terms of a change in form. In principle, such an approach
in bioinformatics are discussed and applied. There are attempts to is more potent than works based on lexical items.
utilise both distance-based and maximum-parsimony models on Nonetheless, most research of the last odd 10 years points
the one hand, and maximum-likelihood models on the other, both into another direction: The pure character-based computation of
combined with and evaluated by statistical models. linguistic data seems insufficient for most tasks, not only in gram-
For example, psychologists, biologists, and evolutionary anthro- matical and comparative linguistic analysis, but also for applied
pologists Gray and Atkinson (2003) published a paper in ‘Nature’ sciences. Spelling checkers, speech recognition, and other forms of
in which a sample from 87 languages was treated to produce lan- automated speech processing require a deeper grammatical under-
guage trees constructed by a ‘restriction site model’ of evolution, standing. The answer to the accumulated problems is only to be
using Markov chains to generate the trees, and autocorrelation found in the development of semi-automated and increasingly
analysis and log-likelihood plots to evaluate the trees (the main automated pre-processing of linguistic data by which grammat-
aim being divergence-time estimations; see for methodical dis- ical information (word-tokens, parts of speech, morphological,
cussion and criticism Holm, 2007). Especially interesting in the syntactic, semantic information, etc.) is fed into the data (cf. the
context of this paper is that the work was based on lexical com- comprehensive discussion in Lüdeling and Kytö, 2008: 484–776, in
parison (‘cognates’), using the so called Swadesh-200 list which particular Fitschen and Gupta, pp.552–564, for morphology).
compiles the lexemes for 200 supposedly culturally independent Seemingly, the nineteenth century resonates in the approaches
concepts, e.g., body parts and basal verbs like eat, drink, see, or hear. which deal with big data-related problems: again, an orientation
It provides a ‘universal’ vocabulary, presumed to be less vulnera- towards the methods of biology can be observed. Language clas-
ble to borrowing (Swadesh, 1952). If for example, the concept Fish sification and comparison again begin with lexical items, but it
is expressed by etymologically related roots (e.g., English fish and becomes increasingly clear that an inclusion of richly processed
German Fisch) the value ‘1’ is entered into the respective matrix morphological and syntactic knowledge is required to improve
cell, if the roots are unrelated ‘0’ (e.g., English bird and French accuracy. In its very own ways, linguistics mirrors the conflict
oiseau for Bird). A binary matrix of 2449 cognates was produced. known from biology between the molecular analysis and the com-
The resulting consensus tree conformed in large parts to estab- parative morphological analysis of its object of research; also in
lished genealogical knowledge but exceeded traditional linguistic the methods it applies (Scholtz, 2010, 2013; Richter and Wirkner,
assumptions of divergence-times. From all methodical issues worth 2014). But it seems that morphological research stands stronger in
discussing, one seems particularly interesting in the context of this linguistics than it does in biology.
paper: since the list only includes uninflected, underivated roots in
their citation form and is only computed for yes/no decisions on 6. Conclusions
cognate state, it does not contain any morphological information
in the grammatical sense of the word. In 2005, Atkinson and Gray Linguistic morphology was ante litteram part of the linguistic
posed a future challenge whose solution might sufficiently improve description since antiquity and is still a vital part of linguistics. It
linguistic comparisons: has fostered and specified the understanding of linguistic kinship
“The sounds comprising each word must be compared across and descent. But morphology has lost grounds in linguistics due to
large sets of data to determine cognacy. Accurate comparisons the focus on syntactic structures and big data accounts.
between words must allow for insertion, deletion, and metathe- This paper has portrayed linguistic morphology as the study of
sis (reversals) and incorporate complex models of phonological the phenomena of accidence and of the properties of form. The
change” (Atkinson and Gray, 2005: 521). interdisciplinary discussion in this volume brings out the relating
facets of morphological research in various disciplines. The com-
In an exploratory work, Hochmuth (2004) used methods for mon root in ancient epistemological philosophy and the repeated
comparing DNA sequences commonly used in bioinformatics joining in theory and methodology support the position that ‘mor-
to achieve this goal. Several versions of the Lord’s Prayer (25 phology’ in these disciplines is not just a shared name. It suggests
diachronic versions from all periods of written records of German; itself that morphology can serve as a universal structural science
21 dialect versions; English, Gothic, Icelandic, Yiddish, Latin, and to a number of disciplines, some of them represented in this vol-
Afrikaans were included for reasons of comparison and evalu- ume. In this regard, biology and linguistics are two sciences strongly
ation) were compared by computing Levenshtein-distances in devoted to morphological research. It should be their objective
substitution matrices. The operations transforming a string of to uphold the comparative analysis of form and function in both
the inner-disciplinary competition for resources and the inter- Li, C.N., Thompson, S.A., 1981. Mandarin Chinese. A Functional Reference
disciplinary commitment to morphological study. Grammar. University of California Press, Berkeley.
Lüdeling, A., Kytö, M. (Eds.), 2008. Handbooks of Linguistics and Communication
Science, vol. 29/1. Corpus linguistics. De Gruyter, Berlin.
References Lyons, J., 1968. Introduction to Theoretical Linguistics. CamUP, Cambridge.
Matthews, P.H., 1974. Morphology. CamUP, Cambridge.
Adelung, J., 1971 [1782]. Umständliches Lehrgebäude der Deutschen Sprache zur McMahon, A., 1994. Understanding Language Change. CamUP, Cambridge.
Erläuterung der Deutschen Sprachlehre für Schulen. 2 vol. Reprint: Olms, Meier, G.F., 1996. Sprachtypologie vor Wilhelm von Humboldt. In: Schmitter, P.
Hildesheim. (Ed.), Sprachtheorien der Neuzeit. Von der ‘Grammaire de Port-Royal’ (1660)
Amundson, R., 2005. The Changing Rule of the Embryo in Evolutionary Biology: zur Konstitution moderner linguistischer Disziplinen, vol. 2. Narr, Tübingen,
Structure and Synthesis. CamUP, New York. pp. 459–477.
Anderson, S.R., 1982. Where’s morphology? Linguist. Inq. 13 (4), 571–612. Messling, M., 2013. La langue du progrès. August Schleicher, l’hégélianisme et
Anderson, S.R., 1992. A-Morphous Morphology. CamUP, Cambridge. l’évolutionnisme linguistique, in: Moussa, S. (Ed.), Le XIXe siècle et ses langues.
Aristotle, 1928. Categories, transl. E.M. Edgehill. In: The works of Aristotle (Ed.), Ve Congrès de la Société d’études romantiques et dix-neuviémistes 2012. Open
vol. 1: Categories and De interpretatione, Ed. by W.D. Ross. Oxford. Access:
Aristotle, 2006. Metaphysics, transl. with an introd. and commentary by Stephen http://etudes-romantiques.ish-lyon.cnrs.fr/wa files/Langues-Messling.pdf
Makin. Clarendon Press, Oxford. Morpurgo-Davies, A., 1998. Nineteenth-Century Linguistics. Longman, London.
Asher, R.E., Simpson, M.Y. (Eds.), 1994. The Encyclopedia of Language and Müller-Wille, St., 2007. Collection and collation. Theory and practice of Linnaean
Linguistics. Pergamon, Oxford. botany. (Studies in History and Philosophy of Science Part C: Studies in History
Atkinson, Q.D., Gray, R.D., 2005. Curious parallels and curious connections. and Philosophy of Biological and Biomedical Sciences 38 (3), pp. 541–562.
Phylogenetic thinking in biology and historical linguistics. Syst. Biol. 54 (4), Osthoff, H., Brugmann, K., 1878–1910. Morphologische Untersuchungen auf dem
513–526. Gebiete der indogermanischen Sprachen, vol. 6. Hirzel, Leipzig.
Bock, W.J., von Wahlert, G., 1965. Adaptation and the form-function complex. Pfeifer, W., 1997. Etymologisches Wörterbuch des Deutschen. Dtv, München.
Evolution 19, 269–299. Plato, 1893. Timaeus, in: The dialogues of Plato, transl. by B. Jowett. Clarendon,
Cassirer, E., 1945. Structuralism in modern linguistics. Word J. Linguist. Circle N. Y. Oxford.
1, 99–120. Ramat, P., 1981. Einführung in das Germanische. Niemeyer, Tübingen.
de Courtenay, J.B., 1972. A baudouin de courtenay anthology. In: Stankiewicz, E. Richter, S., Wirkner, C.S., 2014. A research program for evolutionary morphology. J.
(Ed.), The Beginnings of Structural Linguistics. IndianaUP, Bloomington [1895]. Zool. Syst. Evolutionary Res. 52 (4), 338–350.
Engberg-Pedersen, E., Fortescue, M., Harder, P., Heltoft, L., Falster Jakobsen, L. Robins, R.H., 2000. Classical antiquity. In: Booij. Lehmann, G.E.C., Mugdan, J. (Eds.),
(Eds.), 1996. Content, Expression and Structure. Studies in Danish Functional Handbooks of Linguistics and Communication Science, vol. 17/1. Morphology.
Grammar. Benjamins, Amsterdam. De Gruyter, Berlin, pp. 15–22.
von Goethe, J.W., 1962. Werke, Hamburger Ausgabe. Naturwissenschaftliche Sasse, H.-J., 2001. Scales between nouniness and verbiness. In: Haspelmath, M.,
Schriften I. Allgemeine Naturwissenschaft, Morphologie, Geologie, König, E., Oesterreicher, W., Raible, W. (Eds.), Language Typology and Language
Farbenlehre, Didaktischer Teil, vol. 13. Beck München. Universals. De Gruyter, Berlin, pp. 495–509.
Gray, R.D., Atkinson, Q.D., 2003. Language-tree divergence times support the de Saussure, F., 1983. Course in General Linguistics, transl. and Annotated by R.
Anatolian theory of Indo-European origin. Nature 426, 435–439. Harris Duckworth, London.
Gueintz, C., 1978. Deutscher Sprachlehre Entwurf. Reprint: Olms, Hildesheim Schlegel, F., 1995. Ueber die sprache und weisheit der indier. In: Hutton, C. (Ed.),
[1641]. History of Linguistics. 18th and 19th Century German Linguistics, vol. 4.
Halle, M., Marantz, A., 1994. Some Key Features of Distributed Morphology. Papers Routledge, London [1808].
on Phonology and Morphology, (MIT Working Papers in Linguistics 21), Schleicher, A., 1853. Die ersten Spaltungen des indogermanischen Urvolkes,
Cambridge, 275–288. Allgemeine Monatsschrift für Wissenschaft und Literatur, August 786–787.
Halliday, M.A.K., 1990. An Introduction to Functional Grammar. Arnold, London. Schleicher, A., 1859. Zur Morphologie der Sprache. Eggers, St. Petersburg.
Harris, Z.S., 1946. From morpheme to utterance. Language 22 (3), Schleicher, A., 1861. Compendium der vergleichenden Grammatik der
161–183. indogermanischen Sprachen. Kurzer Abriss der indogermanischen Ursprache,
Hjelmslev, L., 1942. Langue et parole. Cahiers Ferdinand de Saussure publiés par la des Altindischen, Altiranischen, Altgriechischen, Altitalischen, Altkeltischen,
Société Genevoise de Linguistique, 2, 29–44. Altslawischen, Litauischen und Altdeutschen, vol. 2. Böhlau, Weimar.
Hochmuth, M., 2004. Stringbasierte Algorithmen zur Rekonstruktion von Schleicher, A., 1863. Die Darwinsche Theorie und die Sprachwissenschaft. Offenes
Sprachverwandtschaften. Term Paper, Humboldt University, Berlin. Sendschreiben an Herrn Dr. Ernst Haeckel. Böhlau, Weimar.
Hockett, C., 1958. A Course in Modern Linguistics. Macmillan, New York. Scholtz, G., 2010. Deconstructing morphology. Acta Zool. 91, 44–63.
Hockett, C., 1966. Two models of grammatical description. In: Joos, M. (Ed.), Scholtz, G., 2013. Versuch einer analytischen Morphologie. Bildwelten des
Readings in Linguistics, vol. I. University of Chicago Press, Chicago, pp. Wissens, 9, 30–44.
386–400. Schottel, J.G., 1967 [1663]. Ausführliche Arbeit von der Teutschen Hauptsprache.
Holm, H.J., 2007. The new arboretum of Indo-European ‘trees’. Can new algorithms Zilligern, Braunschweig. Reprint: Niemeyer, Tübingen.
reveal the phylogeny and even prehistory of Indo-European? J. Quant. Linguist. Selkirk, E.O., 1982. The Syntax of Words. MIT Press, Cambridge.
3, 167–214. Swadesh, M., 1952. Lexicostatistic dating of prehistoric ethnic contacts. Proc. Am.
Hopper, P.J., Traugott, E., 2003. Grammaticalization. CamUP, Cambridge. Philos. Soc. 96, 452–463.
Hull, D., 2006. Essentialism in taxonomy. Four decades later. In: Wisseman, V. Toman, J., 1998. Word syntax. In: Spencer, A. (Ed.), The Handbook of Morphology.
(Ed.), Annals of the History and Philosophy of Biology 11. Universitätsverlag Blackwell, Oxford, pp. 306–321.
Göttingen, pp. 47–58. Trabant, J., in press. Indien vs. Amerika. Über Friedrich Schlegels Sprache und
Hutton, C., 1995. Introduction. In: Hutton, C. (Ed.), History of Linguistics. 18th and Weisheit der Indier, In: Krämer, Ph., Lenz, M., Messling, M. (Eds.), Rassedenken
19th Century German Linguistics. Friedrich Schlegel. Ueber die Sprache und in der Sprach- und Textreflexion. Kommentierte Grundlagetexte des 19.
Weisheit der Indier, Routledge, London, pp. v–xxx. Jahrhunderts.
Keil, H., 1874. Grammatici Latini vol 6: Scriptores Artis Metricae. Teubner, Leipzig. Wildgen, W., 1984. Goethe als Wegbereiter einer Universalen Morphologie. LAUT,
Kern, P., Zutt, H., 1977. Geschichte des deutschen Flexionssystems. Niemeyer, Trier.
Tübingen. Winford, D., 2013. Contact and borrowing. In: Hickey, R. (Ed.), The Handbook of
Kluge, F., 2002. In: Seebold, E. (Ed.), Etymologisches Wörterbuch. De Gruyter, Language Contact. Blackwell, Oxford, pp. 170–187.
Berlin. Winsor, M.P., 2003. Non-essentialist methods in pre-Darwinian taxonomy. Biol.
Krämer, Ph., 2014. Die französische Kreolistik im 19. Jahrhundert. Rassismus und Philos. 18, 387–400.
Determinismus in der kolonialen Philologie. Buske, Hamburg. Zeige, L.E., 2013. Bildbesprechung: Sprachbilder. Schleichers Sprachstammbaum
Law, V., 2000. The middle ages. In: Booij, G.E., Lehmann, C., Mugdan, J. (Eds.), und Waddingtons Epigenetische Landschaft. Bildwelten des Wissens 9.2,
Handbooks of Linguistics and Communication Science, vol. 17/1. Morphology. 55–60.
De Gruyter, Berlin, pp. 76–90. Ziegler, S., 2000. Das 19. Jahrhundert, in: Booij, G.E., Lehmann, Chr. Mugdan, J.
Law, V., 2003. The History of Linguistics in Europe from Plato to 1600. CamUP, (Eds.), Handbooks of linguistics and communication science, vol. 17/1:
Cambridge. Morphology. De Gruyter, Berlin, pp. 91–103.

1 s2.0 S004452311500008X Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S004452311500008X Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S004452311500008X Main

Uploaded by

Copyright:

Available Formats

Zoologischer Anzeiger 256 (2015) 42–53

Contents lists available at ScienceDirect

Word forms, classiﬁcation, and family trees of languages—Why

Preface itself can be thought of as the abstracted inventory of these

Fig. 2. Early language classiﬁcation (Schottel, 1967 [1663]: 153).

kinship by establishing a new natural science of language which he

You might also like