Folia Linguistica XL/1-2 0165-4004/05/40-51 $ 2.– (C) Mouton de Gruyter, Berlin — Societas Linguistica Europaea
On the Typology of Infl ection Class Systems
Wolfgang U. Dressler, Marianne Kilani-Schoch, Natalia
Gagarina, Lina Pestal, Markus Pöchtrager
Abstract
Infl ectional classes are a property of the ideal infl ecting-fusional language type. Thus strongly infl ecting languages have the most complex vertical and horizontal stratifi cation of hierarchi-cal tree structures. Weakly infl ecting languages which also approach the ideal isolating type or languages which also approach the agglutinating type have much shallower structures. Such properties follow from principles of Natural Morphology and from the distinction of the de-scendent hierarchy of macroclasses, classes, subclasses, subsubclasses etc. and homogeneous microclasses. The main languages of illustration are Latin, Lithuanian, Russian, German, French, Finnish, Hungarian and Turkish.
Introduction
The structure of infl ection class systems is a major object of the third subtheory of Natural Morphology (henceforth NM), viz. the theory of language-specifi c system adequacy, as pioneered by Wurzel (1984; cf. Dressler 2002, Kilani-Schoch & Dressler 2005). In this frame-work infl ection systems of the fol-lowing languages have been dealt with: Bulgarian (Manova & Dressler 2001), Croatian (Dressler, Dziubalska-Kołaczyk & Katičić 1996), Finnish (Pöchtrager et al. 1998), French (Dressler & Kilani-Schoch 2003, Kilani-Schoch & Dressler 2005), German (Dressler 2003), Hungarian (Pöchtrager et al. 1998), Italian (Dressler & Thornton 1996), Latin (Dressler 2002), Polish (Dressler & Dziu-balska-Kołaczyk 1997), Russian (Dressler & Gagarina 1999), Spanish (Aguirre & Dressler, this volume). In this contribution, we will go beyond such accounts of language-specifi c infl ection-class systems and focus on their comparison from the perspective of the second subtheory of NM, viz. the theory of typological adequacy (Dressler 1985, Kilani-Schoch & Dressler 2005). We will try to show the typological relevance of the respective concepts of NM. The languages ana-lysed will be English, Finnish, French, German, Hungarian, Latin, Lithuanian, Russian and Turkish. In general, in other frame-works, infl ection class systems are too vaguely described, not much preciser than Priscian’s description of the traditional Latin infl ection classes (cf. Dressler 2002), at least in respect of the
hierarchy of classes, the differentiation of identical and similar patterning, and the distinction of regular, subregular and irregular paradigms.
Our account starts from the following defi nitions (as exposed in the above descriptions, cf. Kilani-Schoch & Dressler 2005):
Sets of similar infl ectional paradigms form infl ectional classes (in the gener-ic sense), in hierarchgener-ical order: macroclass, class (in the specifi c sense: similar to the traditional term of, e.g., the fi ve Latin declension classes, where not all nouns of one class infl ect in exactly the same way), subclass, (subsubclass, if necessary, etc.), microclass. All classes are defi ned by implicational paradigm structure conditions (cf. Wurzel 1984, Kilani-Schoch & Dressler 2005).
An infl ectional microclass is the smallest subset of an infl ectional class above the paradigm, defi nable as the set of paradigms which share exactly the same morphological generalisations, but may differ via the application of pho-nological processes: phopho-nological assimilation of voicedness in top-s [tøps] vs. dog-s [døgz] does not establish a different plural microclass, whereas morpho-nological assimilation in wive-s [waıvz] does. A paradigm which differs at least in one property from any other paradigm, is an isolated paradigm: these are the only really irregular verbs or nouns (e.g. E. to have, child), whereas small unproductive, minor microclasses are subregular (e.g. E. to sing, goose).
An infl ectional macroclass is the highest, most general type of class, which comprises several (sub-sub-)classes or, at least, microclasses. Prototypically, its nucleus is a productive microclass and it has at least two microclasses. The interior coherence of a macroclass, in terms of shared properties and paradigm structure conditions, must be higher than affi nities between microclasses of dif-ferent macroclasses.
The highest infl ectional set below infl ectional morphology as a whole and above macroclasses is a system of infl ection, i.e. the set of all paradigms formed from bases belonging to the same word class, such as the system of verb infl ec-tion.
The hierarchical make-up of a macroclass follows the principle of default inheritance (cf. Corbett & Fraser 1993) and can be represented by a tree structure, as for German verbs (cf. Dressler 2003) in tree 1.
1. German Verb Infl ection
Macroclass I (weak verbs) comprises all verbs which have a t-preterit and t-past participle (similar to English and Dutch) and the default of no ablaut (apopho-ny). This default is converted into an obligatory property in the fi rst class, e.g. fi rst microclass: spiel-en ‘to play’ – Pret. spiel-te – Past Part. ge-spiel-t, the default microclass, which is the only productive one. In class 2 (microclasses 3 to 5), e.g. microclass 3 (6 verbs), this default is overridden, as in brenn-en ‘to burn’ – brann-te – ge-brann-t. Macroclass II (strong verbs) has at least one
ablaut form (1A: in the preterit), an n-past participle and no t-preterit (unless in transitional forms which are isolated paradigms, such as mahl-en ‘to grind’ – mahl-te – ge-mahl-en, more in Bittner 1996). Moreover macroclass II has the default of at least one umlaut (metaphony: 1U) in the present of its paradigm. Class III overrides this default, class I inherits it (with the exception of micro-class 11), micro-class II converts into just one obligatory umlaut (e.g. schlaf-en ‘to sleep’ – 3.Sg.Prs. schläf-t – schlief – ge-schlaf-en). Furthermore, class I adds a further umlaut form (in the imperative) and in its fi rst subclass a further ablaut to the inherited properties.
2. Infl ecting-Fusional Type
Infl ectional class systems are typical for the infl ecting-fusional language type. Here we defi ne a language type as an ideal construct, to which infl ection systems approach to a greater or smaller degree (according to the typological theory of Skalička 1978; 2002 and its adaptation to the theory of Natural Morphology – cf. Dressler 1985, Kilani-Schoch & Dressler 2005). Thus the ideal infl ecting-fusional type is approached by (especially conservative) Indo-European and Bantu lan-guages, here represented to a large degree (in descending order) by Latin, Lithua-nian and Russian. Languages which mix properties of the infl ecting-fusional and the agglutinating or isolating type, have shallower and more restricted infl ection class tree structures. Among our languages, generally, Hungarian is more agglu-tinating than Finnish, and French is, overall, more isolating than German.
These typological characteristics may not be equal for the infl ection class systems of nouns and verbs within the same language. Thus the French noun sys-tem is much more isolating than its verb syssys-tem (cf. Dressler & Kilani-Schoch 2003, Dressler et al. 2003, Kilani-Schoch & Dressler 2005): the French noun system has only fossile synthetic morphology in terms of rare stem-alternating plural formations of the type chev-al ‘horse’, Pl. chev-aux [ß\'vo], whereas, e.g., English has one productive rule of plural formation plus several marginal and unproductive alternatives (see below) and is thus less isolating than French. In contrast, the French verb system is richer and more complex than the English one (see section 6.).
It holds for an infl ecting-fusional language that the richer and more complex the infl ection system is, the greater is the hierarchical depth and complexity of a macroclass. Moreover stronger infl ecting-fusional languages tend to have more macroclasses and a greater number of productive microclasses. Thus Classical Latin (cf. Dressler 2002) has 4 macroclasses and 6 completely and 5 partially productive microclasses in noun infl ection, three macroclasses and 1 totally and 3 partially productive microclasses in verb infl ection. Russian verb infl ection (Dressler & Gagarina 1999) has 3 macroclasses and 3 totally and 1 partially productive microclasses.
3. Lithuanian Verb Infl ection
The strongest infl ecting-fusional language among contemporary Indo-European languages is Lithuanian with 3 macroclasses and 3 totally productive microclass-es in verb infl ection (out of 44 microclassmicroclass-es), see tremicroclass-es 2a–e: the forms given are 3. Prs., 3. Pret. (number is never differentiated for this person), Inf., 1. Sg. Pres. “A” refers to the accented vowel or diphthong of these four infl ectional forms. The proliferation of microclasses is partially due to morphoprosody, i.e. to morphological distinctions expressed by quantity and pitch contour distinctions of stressed vowels.
4. Russian Verb Infl ection
Russian has three macroclasses. While macroclasses I and II are the biggest classes and contain the majority of microclasses and (monoradical) mini-micro-classes, macroclass III consists of only 8 microclasses (cf. trees 3a–c). The four fully productive microclasses are distributed within the two ‘main’ macroclasses (microclasses 1, 7, 17, 39 – shaded in the tree structure). Despite the long histo-ry of establishing verb classes in Russian, the productive classes in descriptive grammars of Russian (e.g. Belošapkova 1997; Lekant 1982; Miloslavskij 1981) are still numbered arbitrarily after Karcevskij (1927). The number of unproduc-tive groups of verbs, corresponding to microclasses, vary in number from seven (Karcevskij 1927), fi ve (Isačenko 1960), seventeen (Vinogradov 1938/1972), up to twenty (Švedova 1980). Bricyn & Kononenko (1983) claim that “the number of unproductive classes is low and is constantly diminishing”. The present clas-sifi cation (cf. Dressler & Gagarina 1999), however, shows that this is not the case: we assume 50 microclasses. Macroclass I and II differ in that they establish the correlation between the two stems of verbs, the open ‘vowel’ stem and the close ‘consonant’ stem, on the basis of morphotactic addition vs. deletion. Thus, the correlation of the verb stems in macroclass I is Vowel vs. Vowel+Consonant, where the appended consonant is predominantly -j and only marginally -n or -v. Macroclass II is characterized by the deletion of the last vowel in conjunction with various morphonological processes.
5. Latin Verb Infl ection
Morphological richness is established by its productive patterns, complexity also by its unproductive patterns (cf. Dressler 1999). The complex unproduc-tive part of an infl ection system of the infl ecting-fusional type is characterised by the presence of special minor types of unproductive microclasses, so-called mini-microclasses, as can be illustrated with macroclasses I and II of the three macroclasses of Latin verb infl ection (tree 4).
a) mini-microclasses which consist just of 2 or 3 members: e.g. caveo ‘to pay attention’, faveo ’to favour’, Perf. c/favi, Perf.Pass.Part. c/fautum vs. foveo ‘to warm’, moveo ‘to move’, voveo ‘to dedicate’, Perf. f/m/vovi, f/m/votum.
b) monoradical mini-microclasses, such as Lat. video, in-video, pro-video, etc. These verb paradigms cannot be classifi ed as one isolated paradigm, because the meanings of the prefi xed verbs in-video ‘to envy’ and pro-video ‘to provide’ are morphosemantically too opaque for making them compositional derivatives of the simplex video ‘to see’.
c) Some monoradical mini-microclasses may be bound-root microclasses at the same time, e.g. con-sto, con-stare, con-stiti ‘to be certain’ plus in-stare ‘to insist’, ob-stare ‘to withstand’, prae-stare ‘to afford’, etc., vs. the isolated paradigm stare, steti ‘to stand’. Here morphosemantic opacity is also iconically refl ected in more morphotactic opacity in perfect formation.
6. French Verb System
The French verb system presents a typologically interesting case because it equally approaches the infl ecting-fusional and isolating type (cf. Dressler & Kilani-Schoch 2003, Dressler et al. 2003, Kilani-Schoch & Dressler 2005). It has only two macroclasses, such as the weakly infl ecting-fusional languages Ital-ian, Spanish, German, Dutch, English. It has less microclasses than ItalItal-ian, but more than the other languages just cited. It has only two macroclasses, i.e. verbs in Inf. -er (e.g. parl-er ‘speak’ vs. Pres.Sg. [paRl]) on the one hand vs. verbs in -i(-)r (e.g. fi nir ‘fi nish’, sortir ‘go out’ vs. [fi 'ni, søR], -re (e.g. perdre ‘loose’ vs. [p´r]) and -oi(-)r (e.g. croire ‘believe’, pleuvoir ‘rain’ vs. [kRwa, plø]) on the other hand. The above-mentioned weakly infl ecting-fusional languages have two macroclasses as well, in contradistinction to strongly infl ecting languages, which have more macroclasses. However French has less microclasses than Italian, but more than the other languages cited. Thus at least its verb trees are deeper and more complex than those of German (cf. trees 5a–c vs. tree 1).
Surprisingly French has, on the one hand, 3 productive microclasses (types parl-er ‘speak’, sem-er ‘sow’, céd-er ‘yield’), whereas the Germanic languages cited have just 1 (default of weak verbs) and the other Romance languages cited just 1 totally and 1 very weakly productive microclass (as in Spanish, s. Aguirre & Dressler, this volume, for Romanian see Sánchez Miret, this vol-ume). However these 3 productive microclasses of French are phonologically strictly complementary in their root alternations, i.e. in the microclass of parl-er there are no verbs whose stem vowel consists of either a closed or open <e> or schwa followed by a root-fi nal consonant. If a verb has such a stem vowel, it falls automatically either in the second or in the third microclass (sem-er or céd-er). Such a phonological complementarity does not exist in the productive microclasses of the stronger infl ecting-fusional verb systems of Lithuanian and
Russian (see above), Polish (7 productive microclasses), Croatian and Slovene (each 4 productive microclasses). Such phonological complementarity (a dia-chronic consequence of relatively recent morphologization and thus infl ection class split) means greater homogeneity and thus less morphological diversity than is the case with the productive microclasses of the stronger infl ecting-fu-sional languages.
On the other hand, the great majority of the microclasses of the completely unproductive French macroclass II are mini-microclasses or other minor unpro-ductive microclasses (as illustrated with Latin), i.e. 21 of 29 microclasses. Such a relationship does not occur in stronger infl ecting-fusional infl ection systems, but recalls very weakly infl ecting English which abounds in mini-microclasses, such as foot – feet, goose – geese and mouse – mice, louse – lice.
7. Agglutinating Type
A strongly agglutinating language has nearly no infl ection class differentiation at all (cf. Pöchtrager et al. 1998): thus Turkish noun infl ection has just two (produc-tive) microclasses, the default microclass and the phonologically defi ned micro-class of köpek ‘dog’, Acc. köpe-i. Turkish verb infl ection has 3 micromicro-classes: the default one, the verbs with aorist -ir- instead of -er-, and the mini-microclass of ye-mek ‘to say’, de-mek ‘to eat’.
Hungarian noun infl ection is less strongly agglutinating (with a few proper-ties of the infl ecting-fusional type). It has two macroclasses with 3 productive and 13 unproductive microclasses. Its tree structure is very shallow and of little complexity, quite different from the structures of infl ecting-fusional languages, see trees 7a–b (Nom. Sg., Acc. Sg., Poss. Sg.).
Currently a class shift is taking place from the unproductive microclass 13 to the productive microclass 1.
Finnish noun infl ection (where there are still more properties of the infl ect-ing-fusional type than in Hungarian) has two macroclasses and six productive and twelve unproductive microclasses. The tree structure is still shallow, but less so than Hungarian, and has a little more complexity than Hungarian, see trees 7a–b.
In Finnish, the number of productive microclasses is high, but notice the phonological complementarity similar to what we pointed out above for French. The microclass differentiation corresponds in many cases to differing stemfi -nal vowels (in contrast to French, where complementarity is based on the stem vowel) and to only slightly diverging morphonological alternations. The unpro-ductive classes are in most cases associated with a particular derivational suffi x, e.g. microclass 12 contains only nouns with the derivational suffi x -(u)us/-(y)ys ‘-ness, -ity’.
Another characterising property of agglutinating languages, viz. the super-stability of their markers, is not evident from the diagrams, but holds for both
Hungarian and Finnish (as well as evidently for Turkish). Thus, the paradigms resemble each other very much; in Finnish, the little allomorphy which exists in the infl ectional suffi xes can be straightforwardly predicted from the stem shape. As mentioned in the case of French, nominal and verbal infl ection can differ in their complexity. No such imbalance is to be found in Finnish, where the system of noun infl ection has an only slightly more complex hierarchy than verbal infl ection (cf. Pöchtrager et al. 1998).
8. Typological Conclusions
The typological properties identifi ed above can be derived from more basic properties of the respective ideal language type (in the sense of Skalička 1978; 2002 and Dressler 1985): the great preference for biuniqueness of the agglutinat-ing type excludes, in its ideal construct, the differentiation of infl ection classes, whereas the great ambiguity inherent in the ideal infl ecting-fusional type favours it. The absolute morphosemantic transparency of the agglutinating type not only excludes parasitic infl ection (as in the French conditional, which only looks like an imperfect of the future, cf. Kilani-Schoch & Dressler, in print), but also the often opaque category of gender in noun infl ection, whose presence in the infl ecting-fusional type leads to proliferation of infl ection classes in noun mor-phology (the presence of gender in the rather agglutinating Dravidian languages complicates noun infl ection as well). The great preference for morphotactic transparency in agglutinating languages is also an obstacle to the proliferation of infl ection classes, thus the differentiation of infl ection classes in Hungarian and Finnish is due to opacity as well as to partial lack of biuniqueness. In the infl ecting-fusional type, morphotactic opacity (especially in its strong form of fusion) increases class differentiation and makes tree structures deeper and more complex. Finally, in the ideal agglutinating type all morphological patterns are productive, whereas the ideal infl ecting-fusional type has more productive than unproductive infl ectional categories and rules, but due to opacity and ambigu-ity many more unproductive than productive microclasses, which makes tree structures very complex.
Also the ideal isolating type excludes infl ection class differentiation. Thus in the diachronic development, e.g., of Latin to French and of Old to Modern English, certain innovations may superfi cially resemble developments towards an agglutinating language, but profounder analysis (as in Dressler & Kilani-Schoch 2004) shows that the apparent similarity with agglutinating languages is a secondary consequence of greater iconicity shared by weakly infl ecting-fu-sional with agglutinating languages, as opposed to stronger infl ecting languages. But a very close approximation to the ideal isolating type, as in French noun infl ection, reduces iconicity again.
Although infl ection class systems have not yet been investigated in detail for languages approaching the ideal introfl ecting and incorporating types, the criteria introduced in this contribution and derived from higher-order properties of the model of Natural Morphology have shown its potential for typological differentiation for the other morphological language types.
Addresses of the authors
Wolfgang U. Dressler – Lina Pestal – Markus Pöchtrager Institut für Sprachwissenschaft der Universität Wien Berggasse 11 1090 Wien Austria [email protected], [email protected], [email protected] Marianne Kilani-Schoch
Université de Lausanne, Faculté des lettres Ecole de français langue étrangère (FLE)
Quartier UNIL-Dorigny, Bâtiment Humense 2100, 1015 Lausanne
Switzerland
[email protected] Natalia Gagarina
Zentrum für Allgemeine Sprachwissenschaft, Typologie und Universalienforschung (ZAS) Jägerstraße 10 / 11
10117 Berlin Germany
References
Aguirre, Carmen & Wolfgang U. Dressler 2006. “On Spanish verb infl ection.” This volume. Belošapkova, Vera A. 1997. Sovremennyj russkij jazyk. Moskva: Azbukovnik.
Bittner, Andreas. 1996. Starke ‘schwache’ Verben – schwache ‘starke’ Verben: Deutsche Verb-fl exion und Natürlichkeit. Tübingen: Stauffenburg.
Bricyn, Mikhail A. & Viktor I. Kononenko. 1983. Sovremennyj russkij jazyk. Kiev: Višča Škola.
Corbett, Greville & Norman Fraser. 1993. “Network morphology: a DATR account of Russian infl ectional morphology.” Journal of Linguistics 29. 113-142.
Dressler, Wolfgang U. 1985. “Typological aspects of Natural Morphology.” Acta Linguistica Hungarica 35, 51-70.
Dressler, Wolfgang U. 1999. “Ricchezza e complessità morfologica.” In: Fonologia e Morfologia dell’Italiano e dei Dialetti d’Italia. Atti del 21. congresso SLI. Roma: Bulzoni, 587-597. Dressler, Wolfgang U. 2002. “Latin infl ection classes.” In: A.M. Bolkestein et al. eds. Theory
Dressler, Wolfgang U. 2003. “A sketch of dynamic morphology of German verb infl ection.” In: H.G. Jacobsen et al. eds. Take Danish – for instance, Fs. H. Basbøll. Odense: University Press of Southern Denmark, 29-39.
Dressler, Wolfgang U. & Katarzyna Dziubalska-Kołaczyk. 1997. “Polish infl ection classes within Natural Morphology.” Bulletin de la Société Polonaise de Linguistique 53, 95-119. Dressler, Wolfgang U., Katarzyna Dziubalska-Kołaczyk & Antigone Katičić. 1996. “A
contras-tive analysis of verbal infl ection classes in Polish and Croatian.” Suvremena lingvistika 22, 41-42, 127-138.
Dressler, W.U. & Natalia Gagarina. 1999. “Basic questions in establishing the verb classes in contemporary Russian.” In: L. Fleishman et al. eds. Essays in Poetics, Literary History and Linguistics, Fs. V.V. Ivanov. Moskva: OGI, 754-760.
Dressler, W.U. & Marianne Kilani-Schoch. 2003. “Hierarchy and the Classifi cation of French Verbs.” In: M.R. Wise et al. eds. Language and Life: Essays in Memory of K.L. Pike. Dal-las: SIL, 551-567.
Dressler, W.U. & Marianne Kilani-Schoch. 2004. “Iconicité dans la fl exion verbale du français.” In: G. Fenk-Oczlon ed. Sprache und Natürlichkeit, Gedenkschrift für Willi Mayerthaler. Tübingen: Narr, 57-74.
Dressler, W.U., Marianne Kilani-Schoch, Rossella Spina & Anna M. Thornton. 2003. “Le classi di coniugazione in italiano e francese.” In: M. Giacomo-Marcellesi & A. Rocchetti eds. Il Verbo Italiano. Roma: Bulzoni, 397-416.
Dressler, Wolfgang & Jasna Makovec-Černe. 1995. Die ersten Stufen des Erwerbs der slo-wenischen Flexion: Eine Fallstudie. Travaux du Cercle Linguistique de Prague, N.S. 1, 249-260.
Dressler, Wolfgang U. & Anna M. Thornton 1996. “Italian nominal infl ection”. Wiener linguis-tische Gazette 57-59, 1-26.
Isačenko, Aleksandr V. 1960. Grammatičeskij stroj russkogo jazyka v sopostavlenii so slovackim. Bratislava: Slovackaja AN.
Karcevskij, Serge 1927. Système du verbe russe. Essai de linguistique synchronique. Prague. Kilani-Schoch, Marianne & Wolfgang U. Dressler. 2005. Morphologie naturelle et fl exion du
verbe français. Tübingen: Narr.
Kilani-Schoch, Marianne & Wolfgang U. Dressler. in print. “Compositionnalité et iconicité dans la fl exion du verbe français. Une approche de morphologie naturelle.” Verbum.
Lekant, Pawel A. 1982. Sovremennyj russkij literaturnyj jazyk. Moskva: Vysšaja škola. Manova, Stela & W.U.Dressler 2001. “Gender and declensional class in Bulgarian.” Wiener
linguistische Gazette 67-69, 45-81.
Miloslavskij, Igor’ 1981. Morfologičeskije kategorii sovremennogo russkogo jazyka. Moskva: Prosveščenije.
Pöchtrager, Markus, Csanád Bodó, Wolfgang U. Dressler & Teresa Schweiger. 1998. “On some infl ectional properties of the agglutinating type, illustrated from Finnish, Hungarian and Turkish infl ection.” Wiener linguistische Gazette 62-63, 57-92.
Sánchez Miret, Fernando. 2006. “Productivity of the weak verbs in Romanian.” This volume. Skalička, Vladimír. 1978. Typologische Studien. Braunschweig: Vieweg.
Skalička, Vladimír. 2002. Die Typologie des Ungarischen. Travaux du Cercle Linguistique de Prague 4, 101-108.
Švedova, N. J. (ed.) 1980. Russkaja grammatika ... Moskva: Nauka, AN SSSR.
Vinogradov, Viktor V. 1938/1972. Sovremennyj russkij jazyk (grammatičeskoe učenie o slove). Moskva: Vysšaja škola.
Wurzel, Wolfgang U. 1984. Flexionsmorphologie und Natürlichkeit, Berlin: Akademie-Verlag [Engl. translation: Infl ectional Morphology and Naturalness. Dordrecht: Kluwer].
Tree 1: German Verb Infl ection Macr oclass I w eak 2 ha ben 3 br ennen 4 bring en 5 denken Macr oclass II strong Class I 2A Class III 1A Class II 1U , 1A 6 sing en 7 rinnen 8 sterben 9 stehlen 10 br ec hen 11 g ehen 12 sitzen 13 lesen 14 quellen 15 fahr en 16 sc hlaf en 22 kommen 17 bie g en 18 ziehen 19 fl ie ß en 20 b leiben 21 gr eif en 23 stehen Class 1 1 spielen Class 2
Tree 2a: Lithuanian Verb Infl ection
1. Class. Pr et.: -o 1. CS 1. without af fi x 2. with af fi x 1. V identical 1. In fi x 2. Suf fi x 1. A=A=A 2. à -à -Á 2. V : -e-/-i-/-i- 3 . V : → V 1. -n- 2. -s- 1. V identical 2. V : → V 1. À -À -À 4. 6. 7. à -À -À 9. 10. 12. Á -Á -Á 13. V : →V →V sù pa šã˛ la k ˜yla ju ñ ta ka ˜ısta di ñ gsta m ė´ gsta d ū˜ sta sù po šã lo k ìlo j ù to ka ˜ısto di ñ go m ė´ go d ù so sù pti šá lti k ìlti j ù sti ka ˜ısti di ñ gti m ė´ gti d ù sti sup ù šal ù kyl ù junt ù kaist ù dingst ù m ė´ gstu d ū st ù 2. Á /à -Á /à -Á /à 5. 8. à -à -À 11. Vn → V 14. V : →V →V á ug a, d ˜erina ke ˜rta a ñ ka sk ęsta g ę˜sta á ug o , d ˜erino ki ˜rto ã ko skendo g ˜eso á ugti, d ˜erinti ki ˜rsti à kti sk ęsti g èsti á ugu, d ˜erinu kert ù ank ù sk ęstu g ęst ù I. Macr oclass -(i)a 3. Vn → V ga lá nda ga lá ndo ga lą´ sti ga lá ndu 2. Class. Pr et.: -ė 2. OS
2. OS 1. Suf fi x -j- 2. Suf fi x -v-1. -j- in Pres. 2. -j- in Pret. 3. -j- in Pret. & Pres. 1. -uoja- → -avo- 2. -auja- → -avo- 1. A = A = A 2. V: → V 15. 16. Á -Ã -Á 18. Á -Á -Á 20. Á /Ã - 22. 25. dain ú oja bendr á uja g íeda Á /Ã -Á /Ã g y ja p ū˜ va dain ã vo bendr ã vo gied ó jo j ó ja g ìjo p ù vo dain ú oti bendr á uti gied ó ti j ó jo g ýti p ū' ti dain ú oju bendr á uju g íedu j ó ti g yj ù p ū vù j ó ju 1. V → V: 17. Á -Á -Á 19. Ã /À -Á -Á 21. À -À -À 23. À -À -Á 24. Ã -Ã -Ã l ė´ bauja d ù lka g ù ja dal ìja va r dija l ė´ bavo dulk ė´ jo g ù jo dal ìjo va r dijo l ė´ bauti dulk ė´ ti g ù iti dal ýti va r dyti l ė´ bauju dulk ù guj ù dallij ù va r diju
2. Class. Pr et.: -ė 1. -(i)a 2. -a 1. V identical 2. V → V : 3. V alter nation 1. CS 2. OS 1. -e-/-ė -/-e- 2. -e-/-ė -/-ė - 1. V identical 2. V → V 3. V alternation 26. 30. À -Á -À 31. à -Á -Á 32. 34. à -à -à /Á 35. V : → V 37. šau ˜kia sk ìria k ˜e lia gr ˜e sia b ã ra n ˜e ša g ˜e na šau ˜k ė sk ý rė k ė´ lė gr ė˜s ė b ã rė n ˜e šė g ìn ė šau ˜kti sk ìrti k é lti gr ė˜sti b ã rti n è šti gi ñ ti šauki ù skiri ù keli ù gr esi ù bar ù ne šù ge n ù 1. t → č , d → dž 1. Ã=Ã=à 2. Á-Á-Á 27. 28. 29. 33. t → č 36. V → V: 38. ver ˜č ia ke ñ č ia á ud žia pl ˜eč ia p ìla gri ã una ver ˜t ė ke ñ tė á ud ė pl ė˜t ė p ý lė gri õ v ė ver ˜sti k ę˜ sti á usti pl ė˜sti p ìlti gri ã uti ver č iù ken č iù á ud žiu ple č iù pil ù gri ã unu Tree 2c: Lithuanian Verb Infl ection
II. Macr oclass -i 39. Ã /À -Á -Á 40. Á -Á -Á t ù ri s ė´ di tur ė´ jo s ėd ė´ jo tur ė´ ti s éd ė´ ti turi ù s ė´ d žiu III. Macr oclass -o 1. Pr et.: -o 2. Pr et.: -ė 1. A identical 2. Ã /À -Ã /À -Á 41. Ã /À -Á -Á 42. Á -Á -Á 43. Á -Á -Á 44. žì no t ýko r ó do d ã ro žin ó jo t ýkojo r ó d ė d ã rė žin ó ti t ýkoti r ó dyti dar ýti žinau ˜ t ýkau r ó dau dar au Ab br eviations A = accent CS = closed stem OS = open stem
Trees 2d, e: Lithuanian Verb Infl ection
Tree 3a:
RussianVerb Infl ection
Macr oclass I -V (as a par t of a thematic suf fi x) vs. C ad dition (j, n, v) -V vs j- ad dition -V vs. no j-ad dition thematic -V 10. -V vs. 11. -V vs. v-addition n-addition -V in OB -V in OB mc 7 mc 8 mc 9 -a/an, -e/en = -V in CB = -V in CB -ova/uj -eva/uj -va/j mc 1 mc 2 mc 3 mc 4 j mc 5 mc 6 -a/-aj -e/-ej -i/-ij -u/-u -i/ j -y/oj -V is a par t of a pluriphonemic thematic suf fi x with -a
Macr oclass II C-V vs. C (V alter nation with ) CV vs. C without Ca vs. C with stem augmentation stem augmentation C+ them. V C+V is a par t of a pluriphonemic mc 40 mc 41 mc 42 thematic suf fi x -nu Ca vs. Vn ’a vs. Cn ’a vs. Cn jm im fi nal C in OB = fi n. C in the OB alter nates with 38. 39. fi nal C in CB the fi n. C in the CB (v arious past OB past OB mor phonolo gical alter nations) without -nu with -nu -Ca/-C( ’) -Ce/-C( ’) 17.-Ci/ 18. 19.-26. 27.-31. 32.-37. -C ’ -Co/-C( ’) -Ca/-C 1 -Ce/-C 1 -Ci/-C 1 15. 16. no mob . no mob . r .V r. V 12. 13. 14. -ja/-j no mob . mob r .V r. V Tree 3b: RussianVerb Infl ection
Tree 3c: RussianVerb Infl ection Macr oclass III OB ends in consonant Infi nitiv es with fi nal -Ct’ (Cti) Infi nitiv es with fi nal -č’ fi n al C in infi n . ≠ fi nal C in CB (pr esence of -l in the past) fi nal C in infi n . = fi nal C in CB (absence of -l in the past) without V altern. in the past with V altern. in the past 43. without root V altern. in the past masc . -V Ct’ (sti)/-V C 44. with root V altern. in the past masc . -V Cti/-V 1 C 45. -st’/ -d(d’) klast’/kladu (1sg) 46. -Vsti/ -Vd(d’)/ -ol vesti/vedu(1sg)/v’ol 47 . -esti/ -et(t’)/ -ol plesti/pletu (1sg)/pl’ol 49. -č/-g or ž 50. -č/-g or ž with V altern. in the past 48. -č/-k or č with V altern. in the past
I. Macr oclass -a:r e P erfect: -v- -u-laud-ar e PPP: -it- cr ep-ar e sec-ar e II. Macr oclass -e:r e P erfect: -v-/-u- PPP -s P erfect default sigmatic -u- -v- sigmatic asigmatic moneo doceo fl ev i fusion aug eo indulg eo rideo reduplicati v e lengthening monui docui auxi indulsi risi mor deo monit- doct- auct- indult- f o veo faveo monoradical bound-root video possider e
Tree 5a: French Verb Infl ection Macr oclass I Class 1 Class 2 par l-er Pres. 2 bases mc. 1 sub-cl. 1 sub-cl. 2 pay-er sem-er cé d-er mc. 4 mc. 2 mc. 3 Macr oclass II Class 1 Class 2 Class 3 ... inf. -oir ... sub-class 1 sub-class 2 /X-w ar/ /Xw a-r/ ss-cl. 1 ss-cl. 2 cr oi-r e surseoi-r /Xl-war/ /Xv-war/ mc. 17 mc. 18 (-)val-oir de v-oir , rece v-oir mouv-oir mc. 14 mc. 15 mc. 16
Macr oclass II Class 1 Class 2 Class 3 -i ... ... sub-class 1 sub-class 2 no ampli fi cation ampli fi cation ri-r e 2 bases mc. 1 sscl. 1 sscl. 2 simple-base -i ampli fi ed base /i/ ampli fi ed base with def ault sibilant ssscl. 1 ssscl. 2 ssscl. 1 ssscl.2 sibilant amplif. ampl. /v/ /i/ = PP de viant PP écri-r e sssscl. 1 sssscl. 2 mc. 6 sssscl. 1 sssscl. 2 sssscl. 1 sssscl. 2 ampl. /s/ ampl. /z/ Pres. Ind. inter med. PP -u PP -ert 3dpl = sg. = Subj consonant of fr -i-r mc. 13 ssssscl. 1 ssssscl. 2 assaill-i-r cueill-i-r acqu ér -ir sor -t-i-r cour -ir t/vn-ir Simple P ast Simple P ast mc.7 mc. 8 mc. 9 mc. 10 mc. 11 mc. 12 lui-r e PP -t PP -t mc. 3 cui-r e -di-r e mc. 4 mc. 5
Macr oclass II Class 1 Class 2 Class 3 ... ... neither -ir , nor -oir sub-class 1 sub-class 2 inf. = base + C + r inf. = base + r ssc. 1 ssc. 2 ssc. 1 ssc. 2 inf. /dr/ inf. /tr/ w/out ampl. ampl. sssc. 1 sssc. 2 connaî-t-r e me-tt-r e exclu-r e inclu-r e li-r e -fai-r e ampl. cs ampl. /œ/ mc. 24 mc. 25 mc. 26 mc. 27 mc. 28 mc. 29 ren-d-r e absou-d-r e pein-d-r e join-d-r e mc. 19 mc. 21 mc. 22 mc. 23 pr en-d-r e mc. 20
Tree 6a: Hungarian Noun Infl ection Macr oclass I A: B: no V epenthesis with V epenthesis i Acc. mc 1: no V - with V prod. epenthesis epenthesis görl-t in Nom. in Nom. görl-je mc 2 mc 3 mc 4 mc 5 prod. unprod. without with diszk-et hal-at metathesis metathesis diszk-je hal-a bokor teher bokr -ot terh-et
Macr oclass II C: D: E: F: V -lengthening V -shor tening deletion lo w ering mc 6 mc 7 mc 8 no with mc 13 prod. unprod. unprod., unprod., pizza nyár few ation ation few pizza-t nyár -at hid ajtó pizzá-ja nyar -a hid-at ajtó-t hid-ja ajta-ja mc 9 mc 10 mc 1 1 mini-mc 12 unprod. unprod. unprod. unprod. borj-ú apa few hó borj-at apá-t ló hav-at borj-a ap-ja lov-at hav-a lov-a
Macr oclass I stems unifor m: Nom. = G-stem = P-stem Class A Class B
Nom. ends in long v
o
w
el or diphthong
Nom. ends in shor
t v o w e l Microclass 1 Microclass 2 Microclass 3 Microclass 4 Microclass 5 Microclass 6 rising diphthong: others: round V : -i (Pl. → -e-): -a (Pl. → -o-): -a (Pl. → ): suo maa talo pari kana muna Macr oclass II stems distinct Class C Class D Nom. ≠ P-stem ≈ G-stem Nom. ≈ P-stem ≠ G-stem Microclass 7 Subclass C1 Subclass D1 Subclass D2 P-stem =G-stem: Nom. ≠ P-stem ≈ G-stem
G-stem ends in shor
t V G-stem ends in jä rvi long V or V+V Micro- Mini-Micro- Micro- Micro- Subsub- Micro- Micro- Micro- Micro- Micro-class 8 class 9 class 10 class 11 class D1a class 14 class 15 class 16 class 17 class 18 kieli meri -s- ~ Nom. Pl.: -s- ~ -ks-sisar -n- ~ -m-fi nal -e -s- → -t- ~ veri k ä si suppleti v e puhelin kone reng as olut (no har mon y he vonen in par t. sg.) Microclass 12 Microclass 13 -s- ~ -t-jä nis kauneus Trees 7a, b: Finnish Noun Infl ection