Sunteți pe pagina 1din 5

278 Morphological Change, Paradigm Leveling and Analogy

real-time processing, subject to priming effects and other nongrammatical processing.

Bibliography
Bloomfield L (1933). Language. Chicago: University of Chicago Press.

Hale M (2004). Theory and method in historical linguistics. Oxford: Blackwell. Hock H H (1991). Principles of historical linguistics. New York: Mouton de Gruyter.

Morphological Typology
` di Salerno, Salerno, Italy C Iacobini, Universita
2006 Elsevier Ltd. All rights reserved.

At the beginning of the 19th century, linguistic typology established a small set of types, i.e., isolating, agglutinating, fusional (see The Fusional Type, The Agglutinating Type, and The Isolating Type below), to which any single language could be assigned. The main criteria for the assignment were related to word structure. For a long time, word structure remained the dominant criterion to classify languages into types, so that morphological typology is sometimes defined as classical typology as well. Friedrich Schlegel (17721829) and his brother August Wilhelm (17671845), together with Wilhelm von Humboldt (17671835) who was also the first scholar who identified the polysynthetic type (see Polysynthetic Language: Central Siberian Yupik) gave the most important contribution to the assumption that it was possible to describe the whole grammatical structure of a language starting from the way in which relational concepts are morphologically encoded (following suggestions from contemporary research in botany and paleontology). With Edward Sapir (1921) there was an important shift within morphological typology. Abandoning the holistic approach, he underlined the internal inconsistencies of the classical schema of classification, and distinguished and made explicit the relevant parameters for classification. According to Sapir, it is possible (and usual) for the same language to show morphological structures belonging to more than one type. Modern linguistic typology which arose with the work of Joseph Greenberg (1966) attempts to classify languages simultaneously on several dimensions, using implicational types of relations to establish limitations on the range of possible variation occurring within linguistic structures. Therefore, morphology is no longer seen as the most fundamental form of

language classification, but one of the main levels in which languages can be described. In contemporary typology, the classification by morphological types is mostly a convenient way to rapidly identify some morphological characteristic of languages, but it is marginal both in theoretical and descriptive studies. If compared with those of the beginning of the 19th century, the aims of morphological typology have become more modest. However, the role played by morphology in the research of possible patterns of co-variation and correlation with other levels of linguistic analysis is still very relevant today.

Morphological Types
The classification of languages by morphological types is still today part of the standard terminology of linguistics. However, it is also strongly criticized by the majority of typologists for three main reasons: 1. the classification criteria are rather vague and difficult to apply in a consistent way; 2. the morphological type is defined in terms of mutual favorability of properties rather than of implicational correlations, resulting in a low predictive power; 3. morphological typology has a holistic background. Another reason for criticism comes from the emotive appeal of linguistic imperialism: modern linguistics has disavowed the ideological prejudice dating back to the beginning of the 19th century, according to which the fusional morphological type was considered superior to other types both on a functional and on an evolutionary scale (the presumed superiority stemmed from scholars Western-centered standpoint: in fact, all the older and many of the contemporary Indo-European languages can be classified as fusional). There are three fundamental conceptions of language type based on morphological criteria:

Morphological Typology 279

1. the classical (cf. Schlegel, 1808): each type is distinguished from the others in a clear-cut way and is characterized by the presence/absence of one single feature (e.g., languages with or without inflection); 2. the continuum (cf. Sapir, 1921): each type shares with the others a combination of various features, each of them having a continuum of values with two well-defined ends; the structures of a language can be placed on a specific point along the axis anchored by each feature; 3. the ideal (cf. Skalic ka, 1966): each type is an ideal model (which is never fully realized), consisting of a set of features which tend to co-occur. According to the last conception, each morphological type may be described as a combination of functionally interconnected features, which, as a whole, form an ideal construct characterizing (the whole, or some aspects of) the morphology of languages (Sgall, 1995). Languages are rarely pure types; they usually mix elements of different types. Assigning a language to a specific type depends on the preponderance of features considered significant (the quantification of such features is a difficult problem to solve from a practical point of view).

3. how relational concepts are expressed (making a distinction on whether they are expressed through lexical bases or relational elements), and the degree of grammaticalization of relational concepts. With Sapir, there is a move from a taxonomic criterion according to which each language has to be ascribed to one type, to a classifying criterion based on possible types of morphological structures, in the definition of which semantics plays an important role as well. Moreover, it is explicitly recognized that a language can be classified into more than one type, and that types shade one into the other. Greenberg (1954) elaborated Sapirs proposal with the aim of applying it to quantitative evaluations on the morphology of different languages. The three parameters that are nowadays mostly used for the typological classification of languages are the ratio of morphs to word forms, the number of morphemes to morphs, and the degree of wordinternal modification of morphs. The first parameter distinguishes analytic from synthetic languages; the other two distinguish, within the synthetic group, agglutinating from fusional languages.

Lists of Clustering Features Criticism of Morphological Types


The vagueness of the classification in morphological types is shown by the lack of consensus on the number of both types and parameters identifying them. The three main types are: fusional, agglutinating, and isolating. There is no agreement either on whether the polysynthetic type should be considered autonomous or an extreme degree of the agglutinating one (see Polysynthetic Language: Central Siberian Yupik), or on whether Semitic language features are sufficient to distinguish an introflecting type from the fusional one. The classical morphological typology only referred to the formal encoding of single morphological features. Sapir (1921: Chap. 6) recognized the shortcomings and contradictions of the 19th century typologists, and, in modern linguistic terms, made explicit the formal components implicit in previous proposals. He adopted a multidimensional approach to morphological typology that could integrate different dimensions of classification, among which the semantic content of morphological expression as well. He distinguished three main criteria of classification: 1. morphological technique: isolating, agglutinating, fusional, symbolic (internal modification); 2. index of synthesis: analytic, synthetic, and polysynthetic structures; The best-known attempt to establish a list of features that co-occur in morphological types is the one made by Skalic ka (1966), which also includes aspects of word phonology and word order. In the following sections, the features of the three main types (agglutinating, fusional, and isolating) will be given (see also Finnish as an Agglutinating Language; Italian as a Fusional Language; Chinese as an Isolating Language), with the reminder that whereas the final ends of some dimensions can be reached, others are unlikely or impossible to reach. The analysis and discussion of polysynthetic and introflecting types is carried out in sections devoted to them (see Polysynthetic Language: Central Siberian Yupik; Arabic as an Introflecting Language). Suffice it to say here that whereas the polysynthetic type presents a variety of forms and is spread in a lot of languages, the introflecting type is not widely spread and, even in those languages where it is found, it is restricted to a part of morphology. The features that tend to cluster in languages displaying one of the three main morphological types can be listed as shown in the following sections.
The Fusional Type

1. Words are formed by a root and (one or more) inflectional affixes, which are employed as a

280 Morphological Typology

2.

3.

4.

5.

6.

primary means to indicate the grammatical function of the words in the language. Agreement is widely employed. High degree of modification of internal morph boundaries, with a consequently difficult linear segmentation. Tendency to cumulate morphological meanings in a single affix (with consequent asymmetry between the semantic and formal organization of grammatical markers). Word-class distinction is maximal. Inflection is rich, as regards both the number of inflectional classes and the extension of paradigms. Stem suppletion; many cases of both homonymy and synonymy among affixes; clear distinction between inflectional and derivational affixes. A slight correlation with syntax can be seen in the relatively free word order (but there are also fusional languages with a fairly fixed word order).

The Isolating Type

The Agglutinating Type

1. Words are formed by a root and a clearly detachable sequence of affixes, each of them expressing a separate item of meaning. Affixes are widely employed to indicate the relationships between words. Therefore, there are few or no independent relational elements (e.g., pronouns, pre-/postposition, articles, etc.), and a wide use of nominal cases. 2. Very high matching between morphs and morphemes. Morphs are loosely joined together; consequently it is very easy to determine the boundaries between them. 3. Each affix carries only one meaning: no cases of homonymy or synonymy among affixes; the semantic structure is directly reflected in the morphological articulation of the word; no principled limits to the number of affixes in a word. 4. Word-class distinction is minimal: the same affixes tend to occur with roots belonging to different parts of speech (e.g., personal endings to nouns, case endings to verbs); almost the same morphology for adjectives and verbs. No inflectional classes, no gender distinction. 5. Derivational affixes are widely employed in word formation. The distinction between inflectional and derivational affixes is slight. Many affixes reveal their lexical origin to some extent. The latter feature, together with the tendency of affixes to form autonomous syllables and to be relatively unconstrained in number, results in words that are quite long. 6. Relatively fixed word order. Agreement is almost completely absent.

1. Words are monomorphic, invariable, and formed by a single root. Ideally, bound forms are completely missing. Position is the main way of expressing the relationship between independent words. 2. Relational meanings are not overtly expressed, or the same units that normally encode lexical concepts are used for that purpose as separate helper words; the meaning and function of a word considerably depend upon the syntagmatic context. 3. There is little to no morphological complexity. Morphs are clearly identifiable both phonologically and semantically: morph boundaries are sharply defined, phonological form is invariant, there are no instances of overlapping exponence. Derivation is nonexistent, partly replaced by compounding. 4. The distinction in parts of speech is not clear; there is no overt expression of grammatical categorization. 5. Tendency to monosyllabism with no phonetic distinctions between the elements expressing lexical meaning and the ones expressing relational meaning. 6. Rigid word order.
Comparison

The fusional type is differentiated from the isolating type by the use of bound morphs and the clear-cut distinction between word classes; it is differentiated from the agglutinating type by the kind of juncture between morphs, and the nonbiunivocal correspondence between morphs and morphemes. In the synthetic vs. analytic distinction, the fusional and even more the agglutinating type tend toward the synthetic end. Types characterization can be schematically summed up in Table 1.

Contemporary Morphological Typology


In spite of criticism, the classification by morphological types is still convenient and widely used in order to rapidly identify a number of features that tend to co-occur in the morphology of a language. Some authors also use it to assess the extent to which a language moves away from such ideal constructs both in a synchronic and in a diachronic perspective (Dressler, 1985 argues that languages tend to move toward a typological goal according to a linguistic economy criterion; see Natural Morphology). On the contrary, the classification by morphological types is now negligible in contemporary

Morphological Typology 281


Table 1 Features characterizing fusional, agglutinating, and isolating types
1 2 3 4 5 6 7 8 9 10 11 12 13 14

Fusional Agglutinating Isolating

Yes Yes No

No Yes Yes

No Yes Yes

Yes No No

No No Yes

No Yes Yes

No No Yes

Yes Yes No

Yes Yes No

Yes No No

Yes No No

Yes Yes No

No No Yes

Yes No No

(1) affixes; (2) word function determined by position; (3) possibility of segmentation (morph-morpheme correspondence); (4) clear-cut distinction in part-of-speech; (5) tendency to monosyllabism; (6) fixed syntactic order; (7) syntagmatic structures employing particles; (8) noun marked for number; (9) noun marked for case; (10) noun marked for gender; (11) adjectives expressing agreement; (12) synthetic expression of comparison on adjectives; (13) verb: use of analytic structures (vs. affixes); (14) verb: presence of inflectional classes.

typological research, since scholars do not expect the morphological type to correlate in a significant way with other typological parameters. As they refuse the possibility of classifying the whole of a language into a given type, typologists pay increasing attention to the practice of partial typology, focusing on specific areas of linguistic structure (Bynon, 2004). Partial typology analyzes clusters of properties with a view to ascertaining significant connections and hierarchical organization. The main tool for analysis is establishing implication universals according to Greenbergs (1966) suggestions. However, disregarding morphological types does not result in the death of morphological typology: almost half of Greenbergs 45 universals concern morphology. These universals mainly focus on two aspects previously neglected by morphological typology: 1. the relative order in which (derivational and inflectional) concepts are expressed morphologically within word forms; 2. the hierarchy of concepts which a language expresses morphologically. Examples of the first are universals 28 and 39. The former states that if derivational and inflectional affixes are on the same side of the root, then the derivation is always closer to the root. The latter says that in a noun the expression of number is nearly always closer to the base than the expression of the case. Examples of the second are universal 36 If a language has the category of gender, it always has the category of number, and 37 A language never has more gender categories in nonsingular numbers than in the singular. While it is true that only with Greenberg (1966) does syntax start to play a major role in typology, it is to be said that even the famous syntactic universals concerning word order (e.g., universals 3 and 4, which relate VSO order with preposition, and SOV order with postposition) show a link with morphology, for example, via universal 27 (which relates suffixing and postpositional languages on the one hand and prefixing and prepositional languages on the

other). Another example is universal 41, which states the implication between the SOV word order and the presence of a case system. Because of such a link, morphology has acquired importance for scholars interested in word order typology. There are many studies that associate the use of suffixes or prefixes with OV and VO order, respectively (see Cutler et al., 1985; Hawkins and Cutler, 1988; Hawkins and Gilligan, 1988, where affix preference is related with head position and also explained by psycholinguistic factors; see Bybee et al., 1990 for a diachronic approach; Dryer, 1992, who attempts to integrate different approaches). At the interface between morphology and syntax there are works such as Nichols (1986), which distinguishes head- and dependent-marking, and others, such as those dealing with the morphologically based typology of causative constructions. Greenberg (1966) barely takes into account phonology. However, the Universals Archive (cf. Plank and Filimonova) collects some universals concerning the interaction between morphology and phonology. Examples are universal 219 stating that affixes have a more limited inventory of phonemes than roots and universal 713, which correlates agglutinative morphology with vowel harmony, and fusional morphology with stress accent. As far as morphology itself is concerned, the main effect of Greenbergs typology has been to stimulate the investigation of implicational relations between morphological categories, to promote the study of the markedness (overt expression) of morphological values, as well as the investigation of the reasons for the greater or lesser closeness of relational morphemes to the lexical base, and of the role played by morphological heads. One of the main problems that stands before morphological typology is to identify a phenomenon as the same thing in different languages, since the values of one grammatical category may be expressed morphologically in some languages and in a different way in others. Cross-linguistic identification on purely formal criteria is not far-reaching (an example is

282 Morphological Typology

Greenberg universal 26: If a language has discontinuous affixes, it always has either prefixing or suffixing or both); on the other hand, purely functional statements are not sufficient, because, by definition, they need a formal counterpart to be morphological. This type of difficulty explains why studies on inflectional morphology are overwhelmingly prevailing over derivational ones: categories such as case, gender, and number are cross-linguistically better identified and defined (see Blake, 2001; Corbett, 1991, 2000) than derivational ones; furthermore, the internal articulation of inflectional categories, featuring a number of values within one category (e.g., singular, plural for number), favors cross-linguistic comparison within the inflectional domain. The study of the variation in the expression of derivational categories is certainly more difficult and much remains to be done; see Bauer (2002) for a recent attempt.
See also: Arabic as an Introflecting Language; Chinese as an Isolating Language; Finnish as an Agglutinating Language; Italian as a Fusional Language; Natural Morphology; Polysynthetic Language: Central Siberian Yupik.

Bibliography
Bauer L (2002). What you can do with derivational morphology. In Bendjaballah S et al. (eds.) Morphology 2000. Selected papers from the 9th morphology meeting. Amsterdam: Benjamins. Blake B J (2001). Case (2nd edn.). Cambridge: Cambridge University Press. Bybee J L, Pagliuca W & Perkins R D (1990). On the asymmetries in the affixation of grammatical material. In Croft W, Denning K & Kemmer S (eds.) Studies in typology and diachrony. Amsterdam: Benjamins. Bynon T (2004). Approaches to morphological typology. In Booij G, Lehmann C & Mugdan J (eds.) Morphologie morphology, vol. 2. Berlin: De Gruyter.

Corbett G (1991). Gender. Cambridge: Cambridge University Press. Corbett G (2000). Number. Cambridge: Cambridge University Press. Cutler A, Hawkins J A & Gilligan G (1985). The suffixing preference: a processing explanation. Linguistics 23, 723758. Dressler W U (1985). Typological aspects of natural morphology. Acta Linguistica Academiae Scientiarum Hungaricae 35, 5170. Dryer M S (1992). The Greenbergian word order correlations. Language 68, 81138. Greenberg J H (1954). A quantitative approach to the morphological typology of language. In Spencer R F (ed.) Method and perspective in anthropology. Minneapolis: University of Minnesota Press. Greenberg J H (1966/1963). Some universals of grammar with particular reference to the order of meaningful elements. In Greenberg J H (ed.) Universals of language (2nd edn.). Cambridge: MIT Press. Hawkins J A & Cutler A (1988). Psycholinguistic factors in morphological asymmetry. In Hawkins J A (ed.) Explaining Language Universals. Oxford: Blackwell. Hawkins J A & Gilligan G (1988). Prefixing and suffixing universals in relation to basic word order. Lingua 74, 219259. Nichols J (1986). Head-marking, and dependent-marking grammar. Language 62, 56119. Plank F & Filimonova E (eds.) The universals archive, website http://ling.uni-konstanz. de:591/Universals/introduction.html Sapir E (1921). Language: an introduction to the study of speech. New York: Harcourt, Brace & World. Schlegel F (1808). Ueber die Sprache und Weisheit der Indier: Ein Beitrag zur Begru ndung der Altertumskunde. Heidelberg: Mohr and Zimmer. Sgall P (1995). Prague School typology. In Shibatani M & Bynon T (eds.) Approaches to language typology. Oxford: Clarendon Press. Skalic ka V (1966). Ein typologisches Konstrukt. In Skalic ka V (ed.) (1979). Typologische Studien. Braunschweig: Vieweg.

Morphological Universals
S Tatevosov, Moscow State University, Moscow, Russia
2006 Elsevier Ltd. All rights reserved.

Overview
Morphological universals are generalizations about the morphological structure of a possible human language. Languages differ radically as to the properties of their morphological systems, but this variation is

not unlimited: certain morphological patterns are not attested in any language, and morphological universals are explicit statements that impose restrictions on cross-linguistic variation in the domain. Because all theories of morphology (e.g., natural morphology, lexical phonology, a-morphous morphology, distributed morphology, prosodic morphology, to mention only a few) deduce predictions about what morphological properties cannot occur in natural language, any such theory is associated with its own set of

S-ar putea să vă placă și