Dissertation

2023. Guy Tabachnick. Morphological Dependencies.
This dissertation investigates morphological dependencies: correlations between two lexically specific patterns, such as selection of inflectional affixes. Previous work has established that such correlations exist in the lexicon of morphologically rich languages (Ackerman et al., 2009; Wurzel, 1989), but has not systematically tested whether speakers productively extend these patterns to novel words. I present a series of corpus and nonce word studies—in Hungarian, Czech, and Russian—testing whether speakers vary their selection of suffixed forms of novel words based on the forms of that word that are presented to them. In all three cases, speakers vary their responses in accordance with the provided stimuli, demonstrating that they have learned and productively apply morphological dependencies from the lexicon.
I present a theoretical account of morphological dependencies that can account for my experimental results, based on the sublexicon model of phonological learning (Allen & Becker, 2015; Becker & Gouskova, 2016; Gouskova et al., 2015). In this model, speakers index lexically specific behavior with diacritic features attached to underlying forms in lexical entries, and learn generalizations over sublexicons defined as words that share a feature. These generalizations are stored as constraints in phonotactic grammars for each sublexicon, enabling speakers to learn phonological and morphological dependencies predicting words that pattern together. This model provides a unified treatment of morphological dependencies and generalizations that are phonological in nature. My studies show a wide range of learned effects, not limited to those that follow an organizational principle like paradigm uniformity. The sublexicon model assumes that speakers can learn arbitrary generalizations without restrictions, giving it needed flexibility over more restrictive models which rely on notions of morphophonological naturalness.
Click for dissertation

Publications

2022. Alicia Chatten, Kimberley Baxter, Erwanne Mas, Jailyn Peña, Guy Tabachnick, Daniel Duncan, and Laurel MacKenzie. “I’ve always spoke like this, you see”: Preterite-for-participle leveling in American and British Englishes. Forthcoming in American Speech.
Some English verbs use distinct forms for the preterite (e.g. I broke the door) and the past participle (e.g. I’ve broken the door). These verbs may variably show use of the preterite form in place of the participle (e.g. I’ve broke the door), which we call participle leveling. This paper contributes the first detailed variationist study of participle leveling by investigating the phenomenon in perfect constructions using data collected from three corpora of conversational speech: two of American English and one of British English. A striking degree of similarity is found between the three corpora in both the linguistic and the extralinguistic constraints on variation. Constraints on participle leveling include tense of the perfect construction, verb frequency, and phonological similarity between preterite and participle forms. The variable is stable in real time and socially stratified. The paper relates the findings to theoretical linguistic treatments of the variation, and to questions of its origin and spread in Englishes transatlantically.
DOI (open access)

Manuscripts

2023. Hungarian speakers use morphological dependencies in inflecting novel forms. Submitted to Glossa: a journal of general linguistics.
Theories of morphology must account for lexicalized variation: lexical items that differ unpredictably in their inflection must be memorized individually and differ in their stored representation. When tested on such cases, adult speakers usually follow the “law of frequency matching” (Hayes et al. 2009), extending gradient phonological patterns from the lexicon. This paper looks at lexicalized variation in the Hungarian possessive: first, I show that a noun’s choice of possessive is partially predicted by its plural form as well as its phonological shape. Then, using a novel nonce word paradigm, I show that Hungarian speakers productively apply this cooccurrence pattern between the plural and possessive. I treat lexicalized variation as indexed by diacritic features marking lexical entries and propose that Hungarian speakers have learned a gradient cooccurrence relation between diacritic features indexing their plural and possessive forms, extending the sublexicon model of Gouskova et al. (2015). This approach also allows for a flexible analysis of inflection classes in languages like Russian as emergent clusters of frequently cooccurring features.
Click for draft

2022. Intervention effects and phase boundaries in Czech clitic climbing. Submitted to Proceedings for Formal Approaches to Slavic Linguistics (FASL) 31, McMaster University, June 24–26.
This paper examines restrictions on the ability of Czech second-position clitics to climb out of embedded clauses. Clitics cannot climb out of CP phases, but can climb out of infinitives without CP phase boundaries, even those with syntactic subject PRO. Clitics cluster together in a set order, and while arguments of a single verb can freely rearrange themselves to match the required order, an argument of an embedded infinitive generally cannot climb over the controller in the matrix clause in object control constructions. I propose that clitic movement is due to a DP probe that comes equipped with a hierarchy of case features, and that clitics reached by the probe out of order with respect to the hierarchy are trapped and cannot cliticize. Arguments may freely scramble within a single TP, allowing arbitrary reordering, but embedded arguments cannot scramble over matrix arguments, leading to the restrictions in object control sentences.
Click for draft

2021. A Sublexicon Approach to Morphological Dependencies.
Ackerman et al. (2009) and Ackerman and Malouf (2013) argue that the organization of morphological systems allows speakers to efficiently solve the Paradigm Cell Filling Problem: how to predict a word’s inflected form given some of its other forms. For example, a given word’s possessive suffix in Hungarian can be predicted in part by its phonology and by its exponent in forms like the plural. I propose a model in which this phonological and morphological predictability is encoded in formal, constraint-based grammars. In particular, I extend the sublexicon model (Gouskova et al., 2015), in which lexically specific behavior is handled using diacritics on lexical items. Thus, I treat the Paradigm Cell Filling Problem as a problem of finding correlations between diacritics in a given lexical entry, without relying on the storage of a paradigm’s output forms.
Click for draft / lingbuzz

2021. Contextual allomorphy “at a distance” in the Hungarian DP.
Myler (2017) proposes an algorithm for root-outward cyclic Spell-Out (Bobaljik, 2000) where the order in which heads are spelled out may deviate from their linear order when phrasal movement within the phonological word causes a mismatch between this linear order and the syntactic hierarchy (so-called Mirror Principle violations). Myler uses this method to account for cases where a morphophonological process applies across an intervening morpheme. I extend his model to morphosyntactically conditioned contextual allomorphy in Hungarian possessive DPs. Myler’s algorithm allows us to preserve strict locality, where only adjacent morphosyntactic nodes may condition contextual allomorphy (Embick, 2010).
Click for draft / lingbuzz

Presentations

2023. Speakers apply morphological dependencies in the inflection of novel forms. Talk presented at Ling Lunch, University of Connecticut, April 18.
Theories of morphology must account for lexicalized variation: lexical items that differ unpredictably in their inflection must be memorized individually and differ in their stored representation. When tested on such cases, adult speakers usually follow the “law of frequency matching” (Hayes et al. 2009), extending gradient phonological patterns from the lexicon. In this talk, I present results from two wug tests showing that Hungarian and Czech speakers likewise extend gradient morphological patterns from the lexicon: that is, they productively imply correlations between inflected forms of the same word. I handle lexicalized variation using diacritic features marking lexical entries and propose that Hungarian and Czech speakers have learned a gradient cooccurrence relation between diacritic features, extending the sublexicon model of Gouskova et al. (2015). This approach also allows for a flexible analysis of traditional inflection classes (in languages like Russian) as emergent clusters of frequently cooccurring features.
Click for slides

2023. Speakers apply morphological dependencies in the inflection of novel forms. Talk presented at Linguistic Society of America (LSA) 97, Denver, CO, January 5–8.
Since Berko (1958), nonce word studies have shown that speakers exhibit morphological productivity: they can create morphologically complex forms of unfamiliar lexical items. Speakers are known to use a word’s phonology in morphological productivity (e.g. Bybee, 2001; Albright and Hayes, 2003). Using a novel nonce word paradigm, I show that speakers can also be sensitive to morphological behavior: Hungarian speakers take a novel word’s plural into account in selecting its possessive, reflecting the distribution in the lexicon. This experimental paradigm sheds light on how speakers use morphological dependencies: correlations between morphologically complex forms (see Ackerman and Malouf, 2013).
Click for slides

2022. Clitic climbing and negative concord in Czech. Poster presented at Formal Approaches to Slavic Linguistics (FASL) 31, McMaster University, June 24–26.
Czech arguments and reflexives are often expressed as pronominal “second position” clitics that cluster towards the beginning of a clause. Clitics associated with arguments of embedded infinitives can often climb into the cluster position in the matrix clause. If clitic climbing is “clause-bound” (blocked by something like a phase boundary), it should pattern with other clause-bound processes like negative concord, but in fact, clitic climbing is more restricted, being blocked in certain object control constructions. I argue that negative concord and clitic climbing are both blocked by phase boundaries, but that clitic climbing is subject to an additional restraint: it must preserve c-command relations among clitics before movement. To account for the range of clitic climbing, I propose that clitics move to specifiers of dedicated heads and that clitic movement can be preceded by relatively (but not always) free scrambling to the vP edge.
Click for slides

2021. A sublexicon approach to the Paradigm Cell Filling Problem. Talk presented at American International Morphology Meeting (AIMM) 5, Ohio State University, August 26–29.
How do learners figure out an inflected form of a word when they haven’t seen it before and a language allows for more than one option? In some cases, learners can make generalizations about a word’s phonological form (e.g. English verbs ending in [ɪŋ] like sting often have past tenses with [ʌŋ]). In others, as Ackerman et al. (2009) and Ackerman and Malouf (2013) show, knowing some of a word’s inflected forms often allows one to efficiently solve the Paradigm Cell Filling Problem—that is, predicting an additional form. They argue for a morphological model in which the paradigm is a fundamental unit of structure.
I propose a model for how learners may use some forms of a word to predict others outside a paradigm-based formal system. In particular, I extend the sublexicon model (Gouskova et al., 2015; Becker and Gouskova, 2016), used for capturing phonological generalizations, to include dependencies between morphophonological behaviors. This can account for Hungarian possessive allomorphy, in which a noun’s choice of possessive suffix can be substantially, but not entirely, predicted both by its phonological characteristics and its membership in a certain morphological class.
Click for handout

2021. Dependencies between morphophonological patterns. Talk presented at Phonology in the North East (PhoNE), New York University, April 23. Click for handout

2020. Uvular rhotic weakening in Yiddish adjectival suffixes. Poster presented at Acoustics Virtually Everywhere: The 179th Meeting of the Acoustical Society of America (ASA), December 7–11.
In traditional Yiddish dialects, the presence vs. absence of word-final rhotics after [ɜ] in adjectival suffixes carries a heavy functional load, making distinctions of gender, number, and case. Belk et al. (2019) note that some Yiddish speakers with uvular rhotics do not fully articulate them word-finally, endangering this crucial distinction and perhaps contributing to the loss of gender and case in the Yiddish of contemporary Hasidic communities. This study analyzes adjectival endings in publicly available recordings of for speakers with uvular rhotics. The majority of speakers generally do not produce an audible [ʀ] or [ʁ] before consonants, but the rhotic leaves its mark: for some, an underlying rhotic conditions higher F1 on the preceding vowel; others have lowered F2. F1 raising of [ɜ] can also occur when followed by the dorsal fricative [x/χ], suggesting that it is the uvularity of the rhotic that causes F1 raising; F2 lowering is limited to following rhotics, suggesting that this is a rhoticity effect. In addition, vowels followed by underlying coda rhotics are longer in duration. Results indicate that the rhotic triggers phonologized changes in the preceding vowel, while its own realization is weakened, perhaps to an approximant, and masked in the acoustic signal.
Click for poster

2019. Alicia Chatten, Jailyn Peña, Kimberley Baxter, Erwanne Mas, Guy Tabachnick, Daniel Duncan, and Laurel MacKenzie. “I’ve always spoke(n) like this, you see”: Participle leveling in three corpora of English. Poster presented at New Ways of Analyzing Variation (NWAV) 48, University of Oregon, October 10–12.
Some English verbs use distinct forms for the preterite (1) and the past participle (2). These verbs may variably show paradigm leveling, where the preterite form is used in place of the participle (3).
(1) I broke the door. (2) I’ve broken the door. (3) I’ve broke the door.
We contribute the first detailed variationist study of participle leveling by investigating the phenomenon in three corpora: Switchboard, a corpus of 10-minute telephone conversations between American English speakers (Godfrey & Holliman 1997); the Philadelphia Neighborhood Corpus, a corpus of sociolinguistic interviews with Philadelphians (Labov & Rosenfelder 2011); and the Diachronic Electronic Corpus of Tyneside English, a corpus of sociolinguistic interviews with residents of the North East of England (Corrigan et al. 2012). We find a striking degree of similarity between the three corpora in the constraints on variation. The general picture is of socially-evaluated variation affected by both syntactic and paradigmatic factors.
Click for poster

2019. Paradigm Uniformity in Czech Prefix Vocalization. Talk presented at Formal Approaches to Slavic Linguistics (FASL) 28 / American International Morphology Meeting (AIMM) 4, Stony Brook University, May 3–5.
The nature of inflectional paradigms in morphology is controversial, with some (e.g. Bobaljik, 2008) arguing that some supposed paradigmatic effects are instead due to morphosyntactic properties. I look at Czech prefix vocalization, a phenomenon in which consonant-final prefixes sometimes require a vowel (in Czech, this is always [ɛ]) at their end when attaching to a root. I analyze it as morphophonologically driven epenthesis and show that it overapplies across an inflectional paradigm, arguing that the paradigm is a meaningful linguistic unit. I account for prefix vocalization with Optimal Paradigms (McCarthy, 2005).
Click for handout