¡Descarga Critique of Halle's 'Prolegomena' - Understanding Morphemes & Word Formation - Prof. 5508 y más Apuntes en PDF de Lingüística solo en Docsity!
ON HALLE'S "PROLEGOMENA TO A THEORY
OF WORD FORMATION" OR WHAT IS A
LINGUISTIC GENERALIZATION?
HANS U. BOAS
According to Halle (1973: 3) speakers of English know that the adjective trans-form-at-ion-al is composed of the five morphemes shown. This analysis is wrong as far as the morphemic status of -at- and -ion^1 is concerned. Word formation deals with the rules which underlie a speaker's ability to determine the meaning of multimorphemic words which al- ready exist in his language and to form and use new lexical items himself. Both phenomena can only be studied in a nontrivial way if it is assumed that a morpheme is based on a sigmfiant/sigmfie relationship,^2 i.e., that it pairs phonetic-phonological and syntactic with semantic information. Thus a suffix must be marked as to its function, i.e., whether it derives nouns from verbs or adjectives from nouns, etc., and as to its particular shade of meaning. Given these assumptions about word formation and leaving aside the derivational history^3 of the verb transform it is probably impossible that there be a lexical item transformate in the English lan- guage which could be derived by means of a suffix -ate. The suffixes which have this or a similar phonetic make-up and which therefore might have induced Halle to his analysis derive nouns from nouns (type consulate from consul)* adjectives from nouns (type passionate from passion) and verbs from nouns (type hyphenate from hyphen) but do not derive adjectives or nouns from verbs. Deriving a verb transformate from the verb transform is ruled out on quite general grounds. It would run counter to the generalization that English as against some Romance languages or German has no suffixally derived deverbal verbs. This should suffice^5 to show that -at- and -ion in transformational and in many (^1) It is not clear to me whether Halle's wording "... by adding the suffix -at-ion or -ion as 2 transformation and decision" (4) is meant to revise this analysis. Cf. fn. 3. 3 Cf. Marchand, 1969: 1. One might argue that the verb form is derived from the noun form by means of a zero morpheme. See Marchand, 1969. 45 Cf. Marchand, 1969. Non-standard speakers who derive the verb orientate from orientation by back-
Unauthenticated | 81.202.69.
6 HANS U. BOAS
other combinations represent the signifiant of one bound morpheme, namely -ation which enables a speaker of English to derive abstract nouns from verbs. Similar objections apply to Halle's first set of word formation rules (1973: 10) which impose bimorphemic structure on words like tot-al, bro-ther, be-lieve by means of the following rules: [STEM + UT/]A [STEM
- ther]x [be + STEM]v. Such rules can only be justified if as in the case of -at- and -ion semantic considerations are completely disregarded. total, brother and believe are unanalysable units^6 in English, i.e., a speaker has to learn them as unmotivated arbitrary linguistic signs which must only be in accordance with the general phonological constraints of his language. If in Halle's words "a grammar is the formal representa- tion of what a speaker must know about his language" (1973: 1), then it must account for, e.g., the speaker's word formation competence by setting up rule systems which reflect this competence. Since word formation involves the semantic, the syntactic and the phonological components of a grammar one should not for the sake of the simplicity of one component establish generalizations which make counter-in- tuitive predictions^7 about the speaker's over-all competence. Within the framework of his article, which excludes the semantic aspect of word formation, Halle implicitly arrives at the same conclusion when he states that "word formation is a fundamentally different process than phonology" (1973: 15) because its rules have access to different stages in a derivation and apply simultaneously. At another point in his paper Halle claims that derivational and inflec- tional morphology must be handled in a completely parallel fashion (1973:6). This looks like an interesting generalization. But one soon finds out that he reintroduces the distinction through the back door by suggesting that "paradigms must appear as entities in their own right somewhere in a grammar" (1973: 9). Generalizing away the distinction would blur the different roles derivational and inflectional morphemes
formation might get a verb transformate by the same process. But then they are not likely to have the verbs 6 orient and transform in their dialect. As to the notions of analysability and productivity in word formation see Mar- chand, 1969:2ff. 7 A typical example for a generalization which is solely based on phonological evidence and is not corroborated by syntatic or semantic facts is the postulation of a difference in constituent structure between exaltation, relaxation and consultation, information in The Sound Pattern of English (p. 112). The supporting argument given there (fn. 64) is vacuous since the different syntactic behavior of relax and inform under nominalization is predictable from their semantic properties. See my forth- coming dissertation for details.
Unauthenticated | 81.202.69.
8 HANS U. BOAS
permanent memory. It is unlikely that a speaker should memorize hundreds and hundreds of words^11 the meaning and syntactic properties of which may be predicted by the rather general rules of word formation. Apart from that, Halle's suggestion amounts to denying the creative aspect of word formation, which is a strange thing for a generative gram- marian to do. With regard to the problem of relevant linguistic generaliza- tions, 1 think that if generative grammarians would study linguistic semantics^12 and pragmatics as thoroughly as syntax and phonology they would probably discover that the principle of simultaneous rule ap- plication which Halle postulates for word formation has to be extended to all components of a grammar; this would help to find generalizations which reveal correspondences and causal dependencies between the rules of the different components and which thereby would consistently reflect the whole linguistic competence of a speaker.
University of Göttingen
REFERENCES Boas, H. U. 1967 "Adjectival Suffixes in Transpositional and Semantic Function in English" (unpublished paper, University of Tübingen). Forthcoming Adjectivalization in English. Brekle, H. E. 1970 Generative Satzsemantik und transformationeile Syntax im System der englischen Nominalkomposition (München). Chomsky, N. 1972 "Remarks on Nominalization", in: N. Chomsky, Studies on Semantics in Generative Grammar (The Hague). Chomsky, N., and M. Halle 1968 The Sound Pattern of English (New York). Dearmond, C. 1969 "The Concept of Word Derivation", Lingua 22, 329-61. Halle, M. 1973 "Prolegomena to a Theory of Word Formation", Linguistic Inquiry 4, 3-16. Lees, R. B. 1960 The Grammar of English Nominalizations (Bloomington, Indiana). Marchand, H. 1969 The Categories and Types of Present-Day English Word-Formation (München)^2.
(^11) This is restricted to unanalysable linguistic signs and lexicalized combinations. The latter carry additional semantic information which is beyond the predictive power of the rules of word formation. Cf. Marchand, 1969; and Brekle, 1970. 12 This is carried out for nominal compounds in Brekle (1970).
Unauthenticated | 81.202.69.
