2.1 Levelling fairly regular may affect an entire class within a grammatical category operates within the 'paradigm'

Transcription

1 Li7 Historical linguistics Analogy 1 ANALOGY makes related forms more similar to each other in form counteracts sound change: sound change is regular but produces irregularity, analogy is irregular but produces regularity (Sturtevant's Paradox) motivated by psychological grammar-internal factors (contrast sound change) 2 TRADITIONAL TYPOLOGY OF SYSTEMATIC ANALOGY 2.1 Levelling fairly regular may affect an entire class within a grammatical category operates within the 'paradigm' OLD FRENCH 1S aim 'I love' aime 2S aimes aimes 3S aimet aime 1P amons > aimons 2P amez > aimez 3P aiment aimet MODERN FRENCH implies that paradigms have psychological reality for speakers 2.2 Four-part analogy (proportional analogy, analogical extension) generalisation of a morphological operation (e.g. ending) already present in the language operates between lexical items a : a b : b stone : stones [z] cow : cows [z] (instead of kine) the 'a' pattern must be a pattern not just one item residues, especially of frequent words, usually remain (e.g. foot : feet) the generalising pattern must be (synchronically) productive overgeneralisation is comprehensible (mouses vs. hice) applied to new creations or loans this just begs the question (see below) similar analogical creations present in child language (see below) can operate in inflectional (as above) or derivational morphology: land : landscape moon : x = moonscape sea : x = seascape 2.3 Problems Distinguishing the two Sometimes the distinction between the two is unclear. Is this levelling or fourpart analogy? Li7 1 DWW

2 OLD ENGLISH lang lengra lengest MODERN ENGLISH long longer longest The two together? LATIN 'honour' 'sister' nom. sing. honōs > honor soror acc. sing. honōrem sorōrem gen. sing. honōris sorōris dat. sing. honōrī sonōrī honos has honor- in most of its paradigm = levelling there are other words with -r- throughout their paradigm = four-part analogy Which analogy? Less productive patterns sometimes replace more productive patterns e.g. dive : dived > dive : dove (in US English) (on model of drive : drove) sneak : snuck dwarf : dwarfs > dwarf : dwarves (on model of wolf : wolves) scarf : scarves Minor productive patterns sometimes spread by analogy e.g. catch : caught (formerly catched) on model of teach : taught bring : brought : brought > bring : brang : brung (dialectally) on model of sing : sang : sung Very minor patterns sometimes innovate to become regular (Maiden 1998) e.g. Romanian fem. nouns with stressed /a/ form their nom.-acc. plur. and gen.-dat. singular by adding -/i/ and changing /a/ to /ə/ e.g. mare sea, mări seas, barcă boat, bărci boats. The alternation initially arose in two items only (ţară ~ ţări land(s) and pradă ~ prăzi prey, booty ), then spread to (virtually) all feminine nouns. Welsh regular 3s past tense ending -odd appears to have generalised from one verb (lladdodd killed ). The productive form can be low in frequency: the German productive plural ending is -s, but it is low in overall frequency. Conclusions predictability is valued over regularity? regularity applies to morphological subclasses? still leaves the question what makes a process productive? Li7 2 DWW

3 3 TENDENCIES OF ANALOGICAL CHANGE 3.1 Kuryłowicz's laws I. Complex markers replaces simple ones German sing. Gast 'guest' : plur. Gäste (vowel alternation + suffix) hence sing. Baum 'tree' : plur. Baume > Bäume (but cf. long above) II. Basic forms influence derived forms How is basicness determined (for Kuryłowicz 'within their spheres of usage')? Often there is no way of determining what is 'derived', cf. OE cēozan : cēas : gecoren > Mod. English choo[z]e : cho[z]e : cho[z]en OHG kiosan : kōs : gikoran > Mod. German küren : kor : gekoren Backformation is a counterexample (e.g. lazy gives rise to a new verb laze). III. Transparent stem + affix forms are used as a basis for remodelling opaque forms English cow : kine > cow : cows IV. Old (unaffected by analogical change) forms may survive in secondary meanings e.g. brethren, elder., wrought (past participle of work). V. Distinctions of marginal significance may be given up in order to maintain a distinction of central significance e.g. number marking was reestablished in Spanish by giving up case marking: LATIN > EXPECTED SPANISH ACTUAL SPANISH nom. sing. clāvis **llaves llave acc. sing. clāvem llave llave nom. plur. clāvēs llaves llaves acc. plur. clāvēs llaves llaves What is the order of 'significance'? Tense > number > case??? Is it the same for verbs and nouns? Is it the same universally? VI. A native form may be remodelled on the basis of a (more prestigious) non-native form: Welsh sing. pregethwr 'preacher': plur. pregethwrs (dialectally, for standard pregethwyr) on model of English stone : stones 3.2 Maczak's tendencies (Ma czak 1958, 1980) I. Longer words tend to be reformed on the basis of shorter words = Kuryłowicz's II. Li7 3 DWW

4 II. Root alternations tend to be abolished i.e. levelling is more frequent that analogical introduction of alternation = a reformulation of Kuryłowicz's III? And the counterexample to Kuryłowicz's I? III. Longer inflections are remodelled on the basis of shorter ones, except if the shorter suffix is zero e.g. Latin calfacere 'to heat' > calefāre (> Fr. chauffer 'to heat') on model of e.g. amāre 'to love'. IV. Zero endings tend to be replaced by full ones e.g. Old English > Modern English sing. stān : plur. stānas 'stone' sing. word : plur. word 'word' word : words (but the reverse in OE fugol : fuglas > Mod. English fowl : fowl) V. Monosyllabic endings are more frequently replaced by polysyllabic ones than vice versa. Middle Welsh sing. neges : plur. negesseu 'errand(s), message(s)' > ModW. neges : negeseuon (cf. sing. cariad : plur. cariadon 'lover(s)'). VI. The indicative reshapes other moods (subjunctive, imperative) rather than vice versa. VII. The present reshapes other tenses rather than vice versa. VIII. Place names are more archaic than related common nouns (see tendency IX). IX. Place names reform on the basis of local case forms (e.g. dative) e.g. German Baden Baden for earlier nom. Bad : dat. Baden; Latin. aqua 'water' > French eau, but dat. plur. Aquīs > French Aix. 3.3 Superordinate generalisations? (Vincent 1974) I. Basic categories (indicative, present, singular) tend to be the model for analogy II. Clear (long) exponents of grammatical categories tend to spread III. Alternation within a paradigm and multiple expression of the same information tends to be eliminated 4 STUDIES OF ENGLISH STRONG AND IRREGULAR VERBS 4.1 The historical data New members of the string ~ strung class IN OLD ENGLISH fling ~ flung sling ~ slung sting ~ stung string ~ strung hang ~ hung IN DIALECTS ring ~ rung (standard rang) bring ~ brung (standard brought) sneak ~ snuck (US, British sneaked) shake ~ shuck (standard shook) drag ~ drug (standard dragged) Li7 4 DWW

5 stick ~ stuck strike ~ struck dig ~ dug spring ~ sprung (standard sprang) drink ~ drunk (standard drank) sink ~ sunk (standard sank) stink ~ stunk (standard stank) New members of the ride ~ rode class IN OLD ENGLISH dive ~ dove strive ~ strove thrive ~ throve weave ~ wove speak ~ spoke shine ~ shone LATER bide ~ bode drive ~ drove ride ~ rode rise ~ rose smite ~ smote stride ~ strode write ~ wrote break ~ broke freeze ~ froze New members of the break ~ broke class swear ~ swore bear ~ bore wear ~ wore tear ~ tore Other classes have made occasional gains, but have generally lost members. 4.2 Moder (1982) adult subjects asked to form past tense of nonce strong and irregular verbs frequency of regularisation indicates productivity Class Frequency of non-regular- No. of verbs Total frequency isation of nonce verbs (%) in class of class string ~ strung ride ~ rode break ~ broke sing ~ sang feed ~ fed build ~ built find ~ found blow ~ blew catch ~ caught take ~ took slide ~ slid sweep ~ swept spell ~ spelt hit ~ hit bear ~ bore this measure of productivity correlates well with observed changes in the history of English the number of verbs in the class and the frequency of the verbs correlates poorly we re still none the wiser what makes something productive 4.3 Bybee & Slobin (1982) adult and child subjects asked to form the past tense of real verbs (adults under pressure) commonest mistakes : Li7 5 DWW

6 1. inappropriate vowel change in swim ~ swam group 2. regularisation of weep ~ wept group 3. general regularisation of less frequent verbs 4. regularisation of hit ~ hit group 1 3 are common to all ages groups these changes are like historical changes in English 4 accounts for a larger proportion of adult errors than child errors > adults are responsible for change?? READING Anttila, Raimo Analogy. The Hague: Mouton. Bauer, Laurie Morphological productivity. Cambridge: Cambridge University Press. Bybee, Joan L. and Slobin, Dan I Why small children cannot change language on their own: Suggestions from the English past tense. In Papers from the 5th International Conference on Historical Linguistics, edited by A. Ahlqvist. Amsterdam: John Benjamins, Hock, Hans Heinrich Principles of historical linguistics. Berlin: Mouton de Gruyter, chapter 10. Hogg, Richard M Analogy and phonology. Journal of Linguistics 15: Hogg, Richard M Analogy without rule loss. Neophilologus 65: Kiparsky, Paul Remarks on analogical change. In Historical linguistics I, ed. by John M. Anderson and Charles Jones. Amsterdam: North Holland, Kuryłowicz, Jerzy La nature des procès dits 'analogiques'. Acta Linguistica 5: Reprinted in Readings in linguistics II, ed. by E. P. Hamp, F. W. Householder and R. Austerlitz. Chicago: University of Chicago Press, Kuryłowicz, Jerzy The evolution of grammatical categories. Diogenes 51: Lahiri, Aditi Analogy, levelling, markedness: Principles of change in phonology and morphology, Trends in Linguistics. Studies and Monographs 127. Berlin: Mouton de Gruyter. Maiden, Martin Towards an explanation of some morphological changes which 'should never have happened'. In Historical linguistics 1997: Selected papers from the 13th International Conference on Historical Linguistics, Düsseldorf, August 1997, eds. Monika S. Schmid, Jennifer R. Austin and Dieter Stein. Amsterdam: John Benjamins. Mańczak, Witold Tendences générales des changements analogiques. Lingua 7: , Mańczak, Witold Laws of analogy. In Historical morphology, edited by J. Fisiak. The Hague: Mouton, Moder, Carol Lynn Rules and analogy. In Explanation in historical linguistics, eds. Garry W. Davis and Gregory K. Iverson, Amsterdam: John Benjamins. Vincent, Nigel Analogy reconsidered. In Historical linguistics I, ed. by John M. Anderson and Charles Jones. Amsterdam: North Holland, Winters, Margaret E Kuryłowicz, analogical change and cognitive grammar. Cognitive Linguistics 8: Li7 6 DWW