Whole-word frequency effects are shown to exist in what appears to be a completely regular system, the spelling of prefix-final /z/ in Russian. Russian prefixes that underlyingly end in /z/ (roz-, bez-, iz-) end in [s] on the surface when followed by a voiceless consonant. According to the rules of Russian orthography, the surface form, rather than the underlying form, must be reflected in the spelling. However, spelling errors reflecting the underlying form often occur, especially for the prefix bez-. The present paper reports that the error rate, either in natural typing on the web or in a classroom dictation task, for a given word is negatively correlated with the frequency of the word, suggesting that Russian writers rely, to a significant extent, on memory of complete orthographic forms as opposed to the orthographic rule.

 

 

 

Russian velar palatalization changes velars into alveopalatals before certain suffixes, including the stem extension ‑i and the diminutive suffixes ‑ok and ‑ek/ik. While velar palatalization always applies before the relevant suffixes in the established lexicon, it often fails with nonce loanwords before ‑i and ‑ik but not before ‑ok or ‑ek. This is shown to be predicted by the Minimal Generalization Learner (MGL), a model of rule induction and weighting developed by Albright and Hayes (2003), by a novel version of Network Theory (Bybee 1985, 2001), which uses competing unconditional product‑oriented schemas weighted by type frequency and paradigm uniformity constraints, and by stochastic Optimality Theory with language-specific constraints learned using the Gradual Learning Algorithm (GLA, Boersma 1997).

 

 

 

 

One of the most puzzling phenomena in language change is the disappearance of previously productive processes. A morphophonological alternation may be productive at one point in time and lose productivity at a later point. In the historical record, the loss of productivity manifests itself as the accumulation of exceptions, where new words coming into the language are no longer required to obey a constraint that was previously exceptionless. The interesting questions then are: 1) how do the first exceptions appear, and 2) why does the number of exceptions continue to grow as time goes on?

 

 

 

This thesis introduces an experimental paradigm designed to test whether human language learners acquire product-oriented generalizations (e.g., “plurals must end in -i”) and/or source-oriented generalizations (e.g., “add –i to the singular to form the plural”). The paradigm is applied to the morphophonological process of velar palatalization. Ecological validity of the paradigm is confirmed by comparison to corpus data from loanword adaptation in Russian. Characteristics of the training task are shown to influence the extent to which the grammar extracted by a learner is product-oriented or source-oriented. This finding suggests that the shape of the grammar is influenced not only by innate biases of the learner (Universal Grammar) but also characteristics of the learning situation.

 

 

 

 

 

 

Network Theory and Morphological Productivity in Russian. Experimental Evidence from Elicited Production and Ratings. VDM Verlag, 2009, 192 pages

Velar Palatalization in Russian and Artificial Grammar: Constraints on Models of Morphophonology. Laboratory Phonology, 1(2), 361-393.

 

 

 

This book presents an implementation of Joan Bybee''s Network Theory (1985, 2001) in a hybrid model involving representations of both individual tokens and word types, which accounts for effects of word and pattern frequency using the single mechanism of competition for spreading activation. Network Theory is tested against Steven Pinker and colleagues'' Dual Mechanism Model using experimental data examining phonological factors influencing suffix choice with novel words adopted into Russian. Evidence for Network Theory over the Dual Mechanism Model is provided by the findings that all competing suffixes are sensitive to which words the novel word is similar to, with no suffix acting as a default. A dissociation between the defining characteristics of a Dual Mechanism default is reported.

 

 

 

Принято считать, что все русские слова принадлежат той или иной части речи и распределяются по частям речи без остатка [Алпатов 1990]. Случаи, когда одна и та же языковая единица может попасть в несколько групп, можно пересчитать по пальцам, и то многие предпочтут говорить об омонимии или переходе в другую часть речи: наречия и предлоги типа кругом, деепричастия и предлоги типа благодаря, прилагательные и существительные типа столовая и пр.

 

 

 

 

К настоящему моменту в работах по русской морфологии не сложилось единой точки зрения на статус деепричастия как грамматической категории. Кроме того, до сих пор не существует единого описания морфологических свойств деепричастий: стандарта — правил образования — и отклонений от него (данная проблема отмечается также в работе [Добрушина 2009]).

 

 

 

Как известно, предмет и задачи морфонологии понимаются в современной лингвистике неоднозначно, существует множество морфонологических школ и направлений. Но какой бы смысл ни вкладывался в термин „морфонология“, включение в сферу морфонологии изучения нефонетических чередований типа г:ж в берегу – бережешь не вызывает никаких возражений.

 

 

 

Установление парадигматических границ слова составляет одну из наиболее сложных, вместе с тем чрезвычайно актуальных проблем морфологии, поскольку исследователь, занимающийся конкретным морфологическим описанием того или иного языка, весьма часто оказывается перед необходимостью недвусмысленно ответить на вопрос: являются некоторые словоформы формами одного и того же слова или же они представляют разные слова?