Will Oxford

Will Oxford
Visiting Associate Professor, Linguistics, MIT (2022-23)
Associate Professor, Department of Linguistics, University of Manitoba (on leave 2022-23)

Contact will.oxford@umanitoba.ca • oxford@mit.edu • he/him pronouns

Bio I do theoretical and descriptive research on syntax and morphology, usually involving the Algonquian languages and often from a comparative perspective, with an emphasis on agreement and morphosyntactic alignment. My fieldwork experience includes Innu, Ojibwe, Cree, and Oji-Cree (Algonquian) as well as Kapampangan (Philippine). I supervise MA and PhD students in theoretical syntax and Algonquian linguistics.

News I'm teaching at MIT for the 2022-23 academic year. • A paper titled Probe specification and agreement variation: Evidence from the Algonquian inverse will be coming out in Linguistic Inquiry. • A paper titled A tale of two inverses will be coming out in Syntax. • A book titled The Algonquian Inverse is in the works for Oxford University Press.

CV For full information about my research and teaching, as well as complete citations for the papers linked below, see my CV (last update August 2022).

Direct-inverse systems Much of my research has sought to understand the nature of the direct-inverse agreement pattern in Algonquian languages. In 2016-19 I held a SSHRC Insight Development Grant titled Structure and Change in Direct-Inverse Systems with collaborator Heather Bliss.

The Algonquian Inverse (in prep for Oxford University Press). A comprehensive description and analysis of the Algonquian inverse. Chapters: 1 Introduction, 2 Inverse morphology, 3 What is the inverse?, 4 The third-person inverse, 5 The SAP inverse, 6 Formal analysis of the inverse, Appendix A Realization of TA theme signs, Appendix B Verb paradigms.
A tale of two inverses (to appear, Syntax). Inverse agreement in Algonquian languages is argued to correlate with object-over-subject movement in 3-on-3 contexts ('the others see her') but not in 3-on-SAP contexts ('she sees us'). Despite this fundamental syntactic difference, it remains possible to explain why the same inverse morphology appears in the two contexts.
Direct, inverse, and neutral: Refining the description of Algonquian transitive verb forms (to appear, Papers of the 52nd Algonquian Conference, contact me for a copy). Definitions are proposed for the terms "direct" and "inverse" (as they apply to Algonquian) and it is argued that some transitive forms are neither direct nor inverse, but instead "neutral".
Inverse marking and Multiple Agree in Algonquin: Complementarity and variability (2019, Natural Language and Linguistic Theory). Inverse marking and portmanteau agreement are in complementary distribution in Algonquin. The two phenomena are proposed to reflect alternative outcomes of the Agree operation on the Infl head; variation in their patterning follows from variation in the specification of the probe on Infl.
Inverse marking as impoverishment (2017, WCCFL 34 proceedings). The Algonquian inverse marker is proposed to be the elsewhere realization of the object agreement head (Voice). The elsewhere form is realized due to impoverishment whenever the agreement features of Voice are duplicated on Infl.
Variation in TA theme signs (+ references, paradigm handout) (2014, 46th Algonquian Conference presentation). A pan-Algonquian survey of the patterning of "TA theme signs", the set of inflectional suffixes that includes the direct and inverse markers. Variation in the distribution of the inverse marker is shown to form a "staircase" cline that can be described in terms of implicational hierarchies.
When direct and inverse are asymmetrical (Papers of the 49th Algonquian Conference). Direct and inverse forms are often described as being morphologically symmetrical, but in many Algonquian languages, an inverse form shows more agreement morphology than the corresponding direct form. This paper describes the range of variation and seeks a potential explanation.

Agreement and hierarchies The theoretical work below uses Algonquian data to investigate the structural sources of variation in agreement patterns and person hierarchies.

Person restrictions depend on overt agreement, not nominal licensing (to appear, NELS 52 proceedings, with Stefan Keine and Jessica Coon). We argue that the distribution of person restrictions in Algonquian languages, and in particular the interaction of person restrictions with ellipsis, points to an account in which person restrictions reflect feature gluttony (over-agreement) rather than the failure of nominal licensing.
Probe specification and agreement variation: Evidence from the Algonquian inverse (to appear, Linguistic Inquiry). An elaborate crosslinguistic cline in the pan-Algonquian distribution of inverse marking is shown to follow elegantly from a model in which (i) agreement patterns can be determined in the syntactic derivation and (ii) probes have both interaction conditions and satisfaction conditions (Deal 2015).
The Activity Condition as a microparameter (2017, Linguistic Inquiry). When a lower head has agreed with a nominal, can a higher head also agree with that nominal? It has been proposed that the answer is set on a language-by-language basis as a macroparameter, but I show evidence from Algonquian that the answer can vary even within a single language, making this yet another putative macroparameter that turns out to be a microparameter instead.
Probe competition as a source of ergative person splits (2015, NELS 45 proceedings). An earlier analysis of a superset of the data that appeared in the 2017 LI squib. Variations in agreement patterns that can be described under the umbrella of split ergativity are proposed to result from an interaction between two probes.
Size matters: Deriving agreement asymmetries from nominal structure (2017, unpublished manuscript). I show how the "person hierarchy" that conditions certain aspects of Algonquian agreement (SAP > IMPERS > 3.PROX > 3.OBV) can be recast as a hierarchy of nominal size (PhiP < DefP < DP < KP).
Multiple instances of agreement in the clausal spine: Evidence from Algonquian (2014, WCCFL 31 proceedings). The main claim, which I still stand by, is that an Algonquian clause can show more than one instance of "true" agreement determined by the Agree operation. I also suggested that certain layers of Algonquian agreement morphology may result from clitic doubling rather than Agree, but I no longer think this is correct.
Microparameters of agreement: A diachronic perspective on Algonquian verb inflection (2014, University of Toronto dissertation). An analysis of Proto-Algonquian verb inflection and the direct-inverse system, formulated with an eye to crosslinguistic variation and theoretical connections. The analysis of the direct-inverse system has been improved in my subsequent work listed above.

Grammatical relations I'm also interested in the concept of structural subjecthood. What is a subject? How does the determination of subjecthood vary across languages? Are some subjects more "subject-like" than others?

An illusory subject preference in Algonquian agreement (2021, Canadian Journal of Linguistics squib). An apparent preference to index the subject that arises in certain corners of the Algonquian agreement system is argued not to reflect a subject preference at all, but rather a preference to avoid redundantly indexing the object.
Voice-less unergatives: Evidence from Algonquian (2018, Proceedings of WCCFL 35, with Rebecca Tollan). Algonquian unergative subjects differ in their morphosyntax from transitive subjects, a difference that we attribute to the absence of a VoiceP layer in unergatives. This difference entails that unergatives are not concealed transitives, at least not in a fully straightforward sense.
What's the inverse of an inverse? Transitive constructions in Algonquian and Austronesian (CLA 2018 proceedings). The direct/inverse alternation in Algonquian languages and the patient voice/agent voice alternation in Philippine languages are compared and shown to share several specific morphosyntactic properties. Some significant differences between the two systems are also identified.

Realization of inflectional morphology The following papers consider the role of operations such as fission and impoverishment in the determination of inflectional syncretisms and morphological templates.

Fission in Algonquian and the status of morphological templates (2019, WSCLA 23 proceedings). The sequence of affixes in the Algonquian inflectional template parallels the sequence of functional heads in the syntactic structure, but the mapping from syntactic heads to inflectional exponents is not always one-to-one: in several instances, a particular head is realized as a sequence of exponents rather than a single exponent. A morphological operation such as fission can capture this one-to-many relationship between syntax and morphology.
Three Algonquian metasyncretisms (+ notes and references) (2018 MoMOT presentation). Three well-known syncretisms in Algonquian inflection are shown to actually be metasyncretisms, a fact that militates in favour of a morphological impoverishment analysis rather than a deep analysis in which the syncretisms reflect properties of the syntactic structure or a shallow analysis in which the syncretisms reflect the conditioning of particular vocabulary items.
Patterns of syncretism in nominal paradigms: A pan-Algonquian perspective (2017, Papers of the 46th Algonquian Conference, with Heather Bliss). We identify several different syncretisms that apply in the morphological realization of number, gender, and obviation across the Algonquian family.

Language description I have been involved in the documentation and description of Algonquian languages since conducting fieldwork for a dictionary project in the Innu Nation in 2006. My more recent descriptive work takes a pan-Algonquian perspective.

Algonquian languages are not ergative (to appear, Papers of the 52nd Algonquian Conference, with Zlata Odribets; contact me for a copy). The title says it all!
Database of Algonquian Language Structures (work in progress). A project to provide basic information about the sounds and grammar of the Algonquian languages in an easily searchable format. The data includes phonological inventories, verb paradigms, morphological glosses, and historical relationships.
Algonquian (2019, Routledge handbook of North American languages). An overview of the members of the Algonquian family and their phonology, syntax, and morphology.
Algonquian grammar myths (2017, Toronto Working Papers in Linguistics). A discussion of ten tricky points of Algonquian morphosyntax involving direct-inverse marking, person hierarchies, obviation, and transitivity.
A survey of locative expressions in Innu-aimun (2014, Papers of the 42nd Algonquian Conference). Innu-aimun is argued to have three core locative categories: locative nouns, locative demonstratives, and locative prepositions. These three categories form locative phrases, which may be introduced or referred to by two additional locative categories: locative presentatives and locative preverbs. The paper describes these categories, illustrates the structure of locative phrases, and documents some of the syntactic processes that apply to locatives.
Verb morphology on Innu-aimun pronouns: Evidence for the nature of Algonquian wh-questions (2013, Papers of the 41st Algonquian Conference). There is a debate regarding the status of Algonquian wh-questions: are they biclausal (i.e. clefted) or monoclausal? I show that Innu-aimun wh-words inflect for tense and modality like verbs, a fact that weighs in favour of a biclausal analysis. In addition to addressing this debate, the paper also gives a description of Innu-aimun nominal predication and tensed pronouns.
A grammatical study of Innu-aimun particles (2008, Algonquian and Iroquoian Linguistics Memoir 20). A minimally revised version of my 2007 MA thesis.

Phonology These papers take a diachronically-informed perspective on the role of contrast and syllable structure in phonological patterns.

Patterns of contrast in phonological change: Evidence from Algonquian vowel systems (2015, Language). Based on a survey of vowel changes in the Algonquian languages, this article argues that the contrastive structure of the inventory influences the patterning of sound changes. Two inventories that have the same surface shape may nevertheless show systematic differences in the patterning of sound changes due to differences in the organization of contrasts in the inventory. The organization of contrasts can itself change over time, a process that may be called "contrast shift". Some general principles of contrast shift and contrast-sensitive sound change are suggested.
Proto-Algonquian phonotactics (2016, unpublished manuscript). This paper describes the phonotactics of Proto-Algonquian. The syllable template is CGVC. The content of the onset is mostly unrestricted, but the content of the coda is constrained in three ways: place contrasts are neutralized in codas, a coda consonant must be followed by an onset consonant (= coda licensing), and a coda consonant must be more sonorous than the following onset consonant (= syllable contact).
Menominee vowel harmony revisited: A height-based underspecification account (2016, Toronto Working Papers in Linguistics). I argue that the prevailing ATR-based analysis of Menominee vowel harmony entails a distorted view of the Menominee vowel system, which shows no other indications of an ATR contrast. I propose an underspecification analysis in which the conditioning of vowel harmony is expressed in terms of height features, not ATR features.
On the contrastive status of vowel length (2012, presented at the Montreal-Ottawa-Toronto Phonology Workshop). Vowel length contrasts can be seen to fall along a continuum of abstractness, with "pure" duration contrasts at one end and tense-lax contrasts at the other. Using the Modified Contrastive Specification framework (Dresher 2009), I suggest that the steps in this continuum reflect different rankings of the length contrast in the structure of the vowel inventory.

Other Some miscellaneous side-projects, including one on English (!)

From head-final to head-initial in Algonquian: Evidence and implications (2016, Toronto Working Papers in Linguistics). This paper suggests that Algonquian languages show evidence of a long-term shift from head-final to head-initial order in three domains: the PP, the DP, and the clausal spine. Each domain displays traces of an original analytic head-final construction whose head directionality became opaque when the heads became affixes. The new analytic constructions that have subsequently developed in each domain consistently display head-initial order rather than head-final order.
Proximate DP, obviative KP: Balancing the morphosyntax and pragmatics of obviation (2017 CLA proceedings). This paper proposes a theoretical approach to Algonquian obviation that is intended to capture not only its morphosyntactic properties but also the crucial role that discourse factors play in its patterning. The analysis combines two existing ideas: Richards' (2010) proposal that proximate nominals are DP while obviatives are KP and Branigan and MacKenzie's (1999) proposal that the reference of proximate nominals is controlled by an operator in CP.
Variation and change in the Degree Phrase (2017, Linguistic Variation). This squib investigates syntactic variation and change in the Degree Phrase (DegP) using three modifiers in the semantic field of 'different' as a case study: English different and other and French différent. Synchronically, these modifiers show extensive microsyntactic variation, spanning a range of positions from A in the DegP to D in the DP. Diachronically, the items in this class show a tendency to move to higher syntactic positions in a way that is familiar from better-studied syntactic domains.