Greenberg's Universal 28 Revisited

Contents

0  C o n t e n t s  1

  I I n t r o d u c t i o n  2

  II Inflection vs derivation  3

1.  Inflection never changes the syntactic category of a word  3

  derivation may change  it.

2.  IM is relevant to the syntax DM is not  4

3.  Inflection is obligatory derivation is optional  4

4.  Inflection does not change the core lexical meaning  5

  derivation  does

5.  IM is more abstract than DM  5

6.  IM is not restricted by arbitrary limitations DM is  5

  Æ IM is more productive  

7.  IM categories may be expressed cumulatively  6

8.  DM can be semantically irregular IM is semantically regular  6

9.  DM is more relevant to the base than IM  6

10.  DM can be reapplied IM cannot be reapplied  7

11.  Derived lexemes are more likely to be stored  7

  in the lexicon than inflected forms  

12.  Inflection is always peripheral with respect to derivation  7

  III Four theories of seeing inflection and derivation interact  7

1.  The dichotomy approach  7

2.  The SLH (Strong Lexicalist Hypothesis)  8

3.  The continuum approach  8

4.  A tripartition  9

  IV What motivates Universal 28  10

  V Counterexamples to Universal 28  11

1.  Of passers-by and hangers-around  11

2.  Suppletive comparative stems  12

3.  Regular comparative stems  13

4.  Productive cases of inflection followed by derivation  14

5.  Kinderchen and other diminutives of plurals  16

  Yiddish adverbs bislexvayz problematic  

6.  17  17

  Of Äpfelsäfte and heroesisms  

7.  1  7

8.  Georgian preverbs  18

9.  Tagalog moderative verbs  19

  V I C o n c l u s i o n 2  0

  B i b l i o g r a p h y 2  1

II. Introduction

As one of its most important theories, Functional Typology (FT) is a highly empirical approach to all fields of linguistics. Joseph Greenberg of Stanford University, the classic representative of FT, wrote Language Universals, the seminal work that constituted FT, in 1966 1 . FT is called typology because it classifies languages into types according to their features as opposed to generic classifications that group languages according to their language families. It is called functional because it searches for universal tendencies in languages based on the assumption that these tendencies are brought about by the need to fulfill certain functions in communication. FT is thereby the counter approach to Generative Grammar (Functionalism), which assumes that there is a Universal Grammar every human is born with. As a result of trying to find universal categories of linguistic behavior, functional typologists try to extract valid principles by synchronic empirical testing and generalizations. They call these extracted principles universals. One of these universals is Universal 28:

Universal 28: If both the derivation and inflection follow the root, or they both precede the root, the derivatio n is always between the root and the inflection.

According to this Greenbergian universal, we should expect never to see a case in which a root is followed first by an inflectional and then by a derivational suffix. It also excludes cases in which a root is directly preceded by an inflectional prefix which is preceded by a derivational prefix in turn. There appear to be a number of such cases in different languages, however. The goal of this paper is to examine in how far Universal 28 can be saved in view of the cases that have been brought forth in the literature to contradict it. Crucial to this examination will be to first clarify the differences between the central terms ‘inflection’ and ‘derivation’, and how the two interact.

1 Greenberg, Joseph. 1966. Language Universals. The Hague & Paris : Mouton.

III. Inflection vs. derivation

In linguistic textbooks and publications that are not directly concerned with the topic, inflectional morphology (IM) and derivational morphology (DM) are often treated as two clearly and easily distinguishable categories. The boundaries turn out to be a lot fuzzier than this when we take a closer look. As a matter of fact, “the question of how inflection can be distinguished from derivation is one of the classical problems addressed by (..) linguists (TEN

HACKEN 1994: 145) 2 .” BOCHNER (1984: 411) 3 says about inflectional and derivational

morphology that “these concepts are notoriously easier to illustrate than to define” and shies away from a definition that includes more criteria than ANDERSON’S (1982) 4 “inflectional morphology is what is relevant to syntax.” In order to be able to work with Universal 28 at all, our answer to ARONOFF and FUDEMANN’S 5 (2005:160) question “is there really a need for such a distinction?” has to be “yes”. As will be seen, however, there are also approaches that see a cline from derivation to inflection rather than a clear-cut dichotomy between the two. But we will first turn to a broad selection of criteria that have been argued to set apart IM and DM. These criteria will be illustrated by examples, and, where possible, they will be accompanied by counterexamples in order to show that virtually all criteria are relative and not absolute.

1. Inflection never changes the syntactic category of a word, derivation may

change it. (SCALISE 1984: 103 6 , ARONOFF AND FUDEMANN 2005: 162, DRESSLER 1989: 7 7 ) For English this means that when a noun like bed receives the inflectional plural suffix -s it is still a noun, and when a verb like kick receives the 3 rd person singular suffix -s it is still a verb as well. Via DM, verbs can be ‘deverbalized’, however, turning kick into a noun when the derivational suffix -er is added; adjectives can be derived from nouns like geographic from geography and so on. HASPELMATH (2002: 77) 8 claims that this generalization is not correct and he comes up with a number of “inflectional deverbal nouns such as English –ing forms (e.g. my raising (of) this issue).” DRESSLER (1989: 7) also adds in a side note that “participles typically behave like adjectives, infinitives like nouns, both being verb

2 Hacken, P ius ten. 1994. Defining Morpho logy: A Princ ipled Approach to Determining the Boundaries of Compounding, Derivation, and Inflection. Hild esheim: Georg Olms.

3 Bochner, Harry. 1984. Inflection within Derivation. The Linguistic Review 3: 411-421.

4 Anderson, Stephen R 1982. Where’s Morphology? Linguistic Inquiry 13: 571-612.

5 Aronoff, Mark and Kirsten Fudemann. 2005. What is Morphology. Malden: Blackwell.

6 Scalise, Sergio. 1984. Generative Morphology. Dordrecht: Foris.

7 Dressler, Wolfgang. 1989. Prototypical Differences Between Inflection and Derivation. Zeitschrift für Phonetik, Sprachwissenschaft und Kommunikationsforschung 42: 1. 3-10.

8 Haspelmath, Martin. 2002. Understand ing Morphology. London: Arno ld.

inflections.” On the other hand, SCALISE (1984: 103) strengthens the generalization and mentions (unfortunately only briefly and without any examples to support the claim) that there is evidence that DM “always changes the syntactic category of its base, even when the change is not evident.”

2. IM is relevant to the syntax, DM is not.

(HASPELMATH 2002: 70, ARONOFF AND FUDEMANN 2005: 162, DRESSLER 1989: 6, ANDERSON 1982: 609) At this point, HASPELMATH (2002: 70) makes an IM-internal differentiation. He says that the above stated criterion is obviously true only for the inflectional categories of agreement, “because the syntactic relation of agreement is their sole raison d’être.” Thus, in Italian, we have una cosa rivelatrice (‘a revealing thing’) but un gesto rivelatore (‘a revealing gesture’) (DRESSLER 1989: 7). Notice the different inflectional markers on the adjectives in order to agree with the different genders of the nouns. In DRESSLER’S (1989: 7) discussion, concord as a property of IM but not of DM is listed as a separate criterion of distinction. Because of IM’s relevance to syntax, inflected words can also not be replaced by simple forms unlike derived words that can usually be replaced by any simple, underived lexeme that belongs to the same word class.

Again, there are counterexamples: while most case-markers are also prescribed by the context (a direct object must be in the accusative case etc.), categories like tense, aspect or mood “are not obviously relevant to the syntax (HASPELMATH 2002: 72).”

3. Inflection is obligatory; derivation is optional.

(HASPELMATH 2002: 72, , DRESSLER 1989: 6) “In a language with morphological case [like Latin] each noun must appear in a specific case form, but it may be either a derived or a simplex word (DRESSLER 1989: 6).” This way, the lexeme insula can only occur in one of ten case forms, while the base insul by itself is not a possible word (HASPELMATH 2002: 72). HASPELMATH adds, however, that “the application of this criterion is made difficult by the fact that many inflectional paradigms are unlike the

INSULA paradigm in that one of the word-forms bears no affix (or bears a zero-affix) and is

identical with the stem.”

4. Inflection does not change the core lexical meaning, derivation does.

(ARONOFF AND FUDEMANN 2005: 162, HASPELMATH 2002: 73, DRESSLER 1989: 6) Shoot (V) and shoots (V) basically mean the same thing, but a shooting is something altogether different. This criterion reaches its limits when it has to explain the difference in meaning between kind and kindness, for example (HASPELMATH 2002: 73).

5. IM is more abstract than DM.

(HASPELMATH 2002: 73, DRESSLER 1989: 7) The meaning of the derivational –ness in kindness may serve as a counterexample to this criterion as well. But DRESSLER (1989: 7) draws a direct comparison between IM and DM to exemplify the general validity of this criterion: he maintains that inflectional plural is more abstract than derivational collective and that inflectional superlatives such as greatest are more abstract than derivational elatives with prefixes like super- or hyper-. HASPELMATH (2002: 73) calls all inflectional meanings “highly abstract (in some intuitive sense).”

6. IM is not restricted by arbitrary limitations, DM is Æ IM is more productive.

(ARONOFF AND FUDEMANN 2005: 162, HASPELMATH 2002: 75, DRESSLER 1989: 6) The scope of IM is often almost all members of a certain word class. Almost all verbs of English can be inflected for tense, for example, and virtually all nouns take a plural morpheme, in most cases it is even the same suffix –d and –z respectively (or their phonologically conditioned allomorphs). Limitations on inflectional productivity can usually be explained away “by the incompatibility of the inflectional meaning and the base meaning,” as can be seen in examples like *she is knowing me or *mammoths are deader than Neanderthals (HASPELMATH 2002: 75). Derivational morphology is restricted by a number of synchronic and diachronic factors. The difference between primary and secondary affixes is one such factor, and books can be and have been written about other factors. 9 A reason for or an outcome of these (arbitrary) restrictions is great rule variation in DM (cf. –ness vs. –ity), while there is virtual uniformity in IM (English has only about half a dozen inflectional morphemes).

9 e.g. Plag, Ingo. 1999. Morphological Productivity: Structural Constraints in English Derivation. Berlin: Mouton de Gruyter.