Matthew Wolf

Lecturer, Department of Linguistics, Yale University

About me

I'm currently a postdoctoral researcher in the Yale Linguistics Department. I completed my Ph.D. in 2008 at the Linguistics Department at the University of Massachusetts Amherst. Before coming to Yale I was a visiting assistant professor in the Linguistics Department at Georgetown University. See my CV for more details.

My research

I'm interested primarily in phonology, morphology, and their interface. Particular interests in this area include the proper treatment of "process" morphology and of phonologically conditioned allomorphy. More recently, including in my dissertation, I've been working on phonological and morphophonological opacity (particularly forms of opacity which fall outside of the familiar counterfeeding and counterbleeding varieties) and their treatment in serial versions of OT like OT-CC and Harmonic Serialism.

Mutation and other "process" morphology

Wolf, Matthew (2007). For an autosegmental theory of mutation. In Leah Bateman, Michael O'Keefe, Ehren Reilly, and Adam Werle (eds.), University of Massachusetts Occasional Papers in Linguistics 32: Papers in Optimality Theory III. Amherst: GLSA, pp. 315-404. A much shorter version, titled "An autosegmental theory of quirky mutations", is available online in the WCCFL 24 Proceedings.

Argues that morphological feature-changing (mutation/umlaut/ablaut, tonal morphemes, etc.) results from the docking of floating autosegments, and not from constraints that compel alternations directly, like REALIZE-MORPHEME (Kurisu 2001) or anti-faithfulness (Alderete 1999).

Wolf, Matthew (2009). Mutation and learnability in Optimality Theory. In Anisa Schardl, Martin Walkow, and Muhammad Abdurrahman (eds.), Proceedings of the Thirty-Eighth Annual Meeting of the North East Linguistic Society, vol. 2, pp. 469-482.

In this paper I argue that learnability considerations provide a reason to prefer an item-based approach to morphology over a declarative one. The core of the argument rests on languages like Javanese where some marked structure is allowed only when it is created through a morphological mutation process. To learn restrictive grammars of such languages, learners must be biased towards ranking mutation-triggering constraints over faithfulness constraints. I show that, under an item-based, autosegmental theory of mutation, this bias reduces to an instance of an independently required Specific-F >> General-F bias (Smith 2000 et seq.). I also discuss arguments by Adam & Bat-El (2008), based on data from child Hebrew, that learners have the opposite bias, with morpheme-realization constraints starting below faithfulness. I show that, in an item-based theory of affixation, the child Hebrew situation can be understood as an intermediate stage resulting from root-faithfulness being promoted before general faithfulness. Thus, far from being inconsistent, in an item-based theory the Javanese and Hebrew facts reduce to two instances of the same thing.

For a different take on effects like the Javanese one, cast within the theory of Optimal Interleaving proposed in my dissertation, see my PLC 33 paper below.

Kawahara, Shigeto, and Matthew Wolf (2010). On the existence of root-initial-accenting suffixes: An elicitation study of Japanese [-zu]. Linguistics 48.4, pp. 837-864. Prepublication version available here.

Several widely-accepted theories of morphologically-governed accent predict that languages can have pre-accenting suffixes which insert an accent onto the base-initial syllable, as opposed to the more commonly-seen scenario where pre-accenting suffixes add an accent to the base-final syllable. However, this has so far represented a typological gap: no unambiguous cases of root-initial-accenting suffixes have ever been reported. We show that this gap is filled by [-zu], a derivational suffix which has recently emerged in Japanese.

Allomorph selection

Wolf, Matthew (to appear). Lexical insertion occurs in the phonological component. In Bernard Tranel (ed.), Understanding Allomorphy: Perspectives from Optimality Theory. London: Equinox. A prepublication draft is available here.

I propose a late-insertion, OT-based model of morphology in which insertion of vocabulary items occurs in the same module of the grammar as phonological operations. I argue that this makes possible an otherwise-elusive satisfactory account of systems of phonologically-conditioned allomorphy which involve an arbitrary preference among the allomorphs, as well as leading to several other desirable empircal and conceptual consequences. These include the possibility of analyzing phonologically-motivated deponency, permitting an account of listed allomorphy that doesn't require morphemes to have multiple URs, and allowing the possibility of morphemes to be inserted for phonological reasons.

Further development of the material in this paper can be found in chapter 2 of my dissertation and my article to appear in Morphology (see below for both).

Opacity and serial OT

Wolf, Matthew (2008). Optimal Interleaving: Serial Phonology-Morphology Interaction in a Constraint-Based Model. Ph.D. dissertation, University of Massachusetts Amherst.

In my dissertation I propose a model of the phonology-morphology interface, called Optimal Interleaving (OI), which is based on two main premises: that phonology and morphological spell-out occur in the same OT grammar; and that this OT grammar is specifically an OT-CC grammar (McCarthy 2007). Chapters 2 and 3 argue that OI makes desirable predictions about the typology of phonologically-conditioned allomorph selection. Chapter 4 shows that OT-CC can be used to model non-derived environment blocking (NDEB), which under the OI proposal includes blocking in environments that are not morphologically derived. The chapter goes on to show that the OT-CC/OI model of NDEB leads to a number of arguably correct typological predictions. Chapter 5 argues that OT-CC is well-suited two analyzing the other two main kinds of opaque interactions between phonology and morphology: cyclic overapplication (which can be treated as a kind of counterbleeding opacity) and the blocking of phonological processes in morphologically derived environments (which can be treated as a kind of counterfeeding opacity). Chapter 5 also critiques two leading current theories of phonology/morphology opacity, namely Stratal OT and phase theory.

Wolf, Matthew (2009). Local ordering in phonology-morphology interleaving: Evidence for OT-CC. Talk presented at 83rd Linguistic Society of America Annual Meeting, San Francisco. Handout

Buckley (1994) shows that the interaction of word-final vowel epenthesis with the selection of allomorphs of the plural and possessive suffixes in Tigrinya creates an ordering paradox in Lexical Phonology. I show that the OT-CC based theory of phonology-morphology interleaving proposed in my dissertation can handle these facts. This result exemplifies the ability of OT-CC to model at least some types of local odering effects (Anderson 1969, 1974).

Wolf, Matthew (2010). On the existence of counterfeeding from the past. Talk presented at 84th Linguistic Society of America Annual Meeting, Baltimore. Corrected handout

Wilson (2006) shows that OT-CC (McCarthy 2007) can produce a heretofore laregely unrecognized mode of opaque process interaction which he dubs "counterfeeding from the past". This talk discusses the prediction, and presents a number of examples from the literature of opaque interactions which seem to be of the counterfeeding from the past variety. This suggests that the prediction pointed out by Wilson may represent a virtue rather than a liability of OT-CC. It also means that OT-CC may be preferable to traditional rule-based phonology as a theory of opacity, since counterfeeding-from-the-past scenarios involve paradoxical rule orderings, which is the point which most of the examples discussed were originally raised in the literature to make.

Wolf, Matthew (2010). Implications of affix-protecting junctural underapplication. In Jon Scott Stevens (ed.), University of Pennsylvania Working Papers in Linguistics 16.1: Proceedings of the 33rd Annual Penn Linguistics Colloquium. Philadelphia: Penn Linguistics Club. Proceedings version; Prepublication draft (proceedings version has some pagination issues, so the draft may be more readable)

Optimal Interleaving theory treats the underapplication of phonological processes in morphologically derived environments as a form of counterfeeding opacity. OI therefore makes more lenient typological predictions than either base-identity constraints (Benua 1997) or phase-based approaches to cyclicity (Marvin 2002). Specifically, OI predicts there to be cases of junctural underapplication in which affix material, rather than base material, is protected against change. This paper argues that this prediction is correct. Examples of affix-protecting underapplication from a number of languages are presented.

Wolf, Matthew (2011). Limits on global rules in Optimality Theory with Candidate Chains. Phonology 28.1, pp. 87-128.

Prepublication draft available here. Please cite from the published version (email me if you don't have access to the journal).

In this article I investigate two types of opaque interaction which are predicted under simultaneous rule application but not under serial rule application: mutual counterfeeding and mutual counterbleeding, in addition to self-counterfeeding, inquiring whether they can be modeled using OT-CC. I show that none can be under the original version of the theory, and demonstrate simple modifications to that version which will make them possible. These results are important because they establish that OT-CC's capacity for modeling global rules (Lakoff 1970) is not unlimited.

Wolf, Matthew (to appear). Cyclicity and non-cyclicity in Maltese: Local ordering of phonology and morphology in OT-CC. To appear in a volume on Harmonic Serialism and Harmonic Grammar edited by John McCarthy and Joe Pater. Current draft available on LingBuzz.

An earlier version of this was presented at the 3rd Conference on Maltese Linguistics (Valletta, April 2011). The handout is available here.

Maltese verbs with pronominal (object-marking) suffixes famously show evidence of cyclic stress, in the the form of underapplication of syncope (Brame 1974). But vowel-final (third weak radical) stems fail to show evidence of cyclic stress (Odden 1993). This talk presents an OI analysis of the difference between C-final and V-final stems, relying on OI's capacity for modeling local-ordering effects.

Wolf, Matthew (to appear). Candidate chains, unfaithful spell-out, and outwards-looking phonologically-conditioned allomorphy. Invited contribution to special issue of Morphology on “New theoretical tools in the modeling of morphological exponence”, guest-edited by Jochen Trommer. Prepublication draft available on LingBuzz.

In parallel OT where the form of entire utterances is computed at once, phonologically-conditioned suppletion of a given morpheme may be sensitive to the phonology of any structurally "outer" material. By contrast, in a cyclic and serial framework, phonologically-conditioned suppletion can only ever be inward-looking, i.e. it may only look at the phonology of what has been spelled out already. OI recaptitulates this sort of local computation by requiring that all alternative morph-insertions for the same morpheme compete for Local Optimality. This paper suggests a relaxation of this assumption, specifically that different morph-insertions compete for Local Optimality only if they have the same (un)faithfulness status with respect to morphosyntactic structure. This assumption brings morph-insertion more closely in line with OT-CC's assumptions about phonological operations being classified according to the basic faithfulness constraint they violate, and it results in the empirical prediction that outwards-sensitive phonologically-conditioned suppletion will be limited to cases where one of the competing allomorphs is morphosyntactically anomalous in some way, e.g. is specified for a feature not present in the input. This paper argues that this prediction correctly characterizes the cases where outwards sensitivity is known to occur.