Papers

2022. Ash Asudeh & Daniel Siddiqi. Morphology in LFG. In Mary Dalrymple, ed., The Handbook of Lexical Functional Grammar. Berlin: Language Science Press. To appear. Unreviewed draft: Please do not cite or quote without permission.

Abstract. Lexical-Functional Grammar has been consistent over the past four+ decades about its conception of syntactic structure and the sorts of rules that license it. However, despite being a highly lexicalist model of grammar, LFG has not developed a similarly consistent model of word-formation. LFG has in fact assumed a variety of different models of word-formation and interfaces with distinct ‘morphological’ modules and theories in this time. This is perhaps because LFG early on solved the problem of how morphology and syntax can communicate in a common formal language — the language of functional descriptions, which can be both associated with words and their parts and with syntactic elements. We first introduce some important concepts from morphological theory. We then look at some early LFG analyses which treated morphology incrementally. Next we review work on the syntax–morphology interface in LFG, which sets the stage for current approaches to morphology in LFG, which are realizational.


2021. Ash Asudeh, Paul B. Melchin & Daniel Siddiqi. Constraints all the way down: DM in a representational model of grammar. In WCCFL 39 Proceedings. Somerville, MA: Cascadilla Press. To appear.

Abstract. Distributed Morphology (DM) is a realizational framework for morphological theory and, like any such framework, it assumes an interface with a syntactic module that provides the structures to be realized. However, unlike other realizational models (e.g., Paradigm Function Morphology), DM has assumed a particular syntactic framework for providing those structures: standard Chomskyan syntax (here called Minimalism for short, even though initial work in DM assumed a version of syntax that was antecedent to Minimalism). Given this consistent co-occurrence, it can be hard to easily identify where Minimalism ends and where DM begins. For example, the Y-model is a feature of Minimalism, as is the fact that head-movement and lowering are in the PF branch of that Y. In this paper, we identify the key features of DM that are independent of its interface with Minimalism and identify those features as effectively comprising DM in isolation.

We then offer our framework of Lexical-Realizational Functional Grammar (LRFG), which marries DM as a theory of morphological realization to Lexical-Functional Grammar (LFG) as a theory of syntax and grammatical architecture. We show that the result keeps all the key features of DM and thus is properly a daughter of DM, offering the same fundamentals of DM but also having different properties from DM’s traditional form, thus allowing it to provide different types of explanations. We demonstrate LRFG by providing a sketch of Ojibwe agreement morphology as a case study (adapting insights from prior analyses). We close by summarizing potential reasons to adopt LRFG as a version of DM.


2021. Daniel Siddiqi. On the Taxonomy of Root Suppletion. In WCCFL 39 Proceedings. Somerville, MA: Cascadilla Press. To appear.

Abstract. This paper concerns a longstanding debate that is prevalent in lexical-realizational models of morphology (to use Stump 2001’s typology)—especially Distributed Morphology, Exoskeletal Syntax, Nanosyntax, and Lexical-Realizational Functional Grammar. The broader debate hinges on the question of how Roots are individuated. Abstract, semantic, and phonological individuation are the competing hypotheses for the individuation of Roots.

The paper does not provide arguments for or against the semantic or abstract individuation of Roots, nor does it offer new arguments against the Phonological Individuation Hypothesis (PIH). Rather, it returns to an empirical argument made throughout the literature that the existence of Root suppletion falsifies the PIH. This argument echoes an argument made by Marantz: that the existence of Root suppletion would mandate some form of Root individuation. This argument is simple: If Root suppletion exists, then Roots do not have consistent phonology, and thus phonology cannot be what identifies or individuates them. Because this is an empirical argument, it hinges on agreed-upon standards of what would constitute empirical counter-evidence. Putative examples of Root suppletion abound, so this standard must delineate what constitutes what the literature has called “true” Root suppletion. In other words, this debate has become taxonomical.

This paper proposes such a standard—one by which we can determine whether a putatively suppletive phonology-semantics covariance constitutes counterevidence to the PIH. This standard synthesizes classical arguments on the nature of suppletion as well as contemporary arguments from Borer and Marantz about what characteristics of a covariance would unambiguously be “true” Root suppletion.

Once proposed, the standard is tested against familiar suppletion data. In the end, the standard does indeed exclude much of the data we are familiar with, but the Root suppletion in the Ojibwe verb amw-miiji ‘eat’, which suppletes for argument animacy, survives the standard, as does the Hopi nominal suppletion, which suppletes for number (e.g. wùuti ‘woman.SG’ / wùuti.t ‘woman.DL’ / momoya.m ‘woman.PL’). These alternations then do, unambiguously, provide counter-evidence to the PIH, falsifying it.


2021. Michael Everdell & Paul B. Melchin. Control the sentence, subordinate the pronoun: on the status of controlled versus non-controlled complement clauses in O’dam. In WCCFL 39 Proceedings. Somerville, MA: Cascadilla Press. To appear.

Abstract. This paper examines the argumenthood properties of two types of clausal complements in the Uto-Aztecan language O’dam (glottocode: sout2976), Controlled and Non-Controlled complement clauses. Using the LFG framework we will argue that Controlled complements are category COMP (i.e. clausal in nature), while Non-Controlled complements are category OBJ (i.e. nominal in nature), following Dalrymple & Lødrup. The referent of the nominal Non-Controlled complement is the eventuality of the Non-Controlled complement’s verb, while its dependents are introduced via a headless relative clause.


2021. Michael Everdell, Paul B. Melchin, Ash Asudeh & Daniel Siddiqi. Beyond c-structure and f-structure: On the argument-adjunct distinction in O’dam. In Miriam Butt and Jamie Y. Findlay and Ida Toivonen, eds., Proceedings of the LFG21 Conference. Stanford, CA: CSLI Publications. 125-145.

Abstract. In this paper we examine the argumenthood properties of Controlled Complement Clauses and Non-Complement Subordinate Clauses in O’dam. We show that in O’dam only controlled COMPs are arguments, while other putative complement clauses are adjunct relative clauses that elaborate on a pronominal OBJ incorporated in the matrix verb. We use the LRFG framework to capture both the argumenthood properties of the two types of clauses in O’dam as well as the patterns of object marking on the matrix verb by taking advantage of mismatches between c-structure (phrase structure and f-descriptions) and v-structure (the vocabulary items realizing this structure).


2020. Paul B. Melchin, Ash Asudeh & Dan Siddiqi. Ojibwe agreement in Lexical-Realizational Functional Grammar. In Miriam Butt and Ida Toivonen, eds., Proceedings of the LFG20 Conference. Stanford, CA: CSLI Publications. 268–288.

Abstract. Lexical-Realizational Functional Grammar (LRFG) is a novel theoretical framework that incorporates the realizational, morpheme-based approach to word-formation of Distributed Morphology into the declarative, modular framework of LFG. LRFG differs from standard LFG in that terminal nodes of c-structure are not words, but are bundles of features that are realized in a separate, linearized v-structure. The mapping from cto v-structure is many-to-one, using the mechanism of Spanning. In this paper we demonstrate LRFG with an account of a part of the Ojibwe (Algonquian) verbal agreement system. We provide descriptions of the relevant templates and vocabulary items and discussion of some relevant examples.


2020. Paul B. Melchin, Ash Asudeh & Dan Siddiqi. Ojibwe Agreement in a representational, morphemebased framework. In M. Emma Butterworth and Angelica Hernandez, eds., Proceedings of the Canadian Linguistic Association. No pagination.

Abstract. We have been developing a theoretical framework that couples Lexical-Functional Grammar (LFG) with the realizational, morpheme-based approach to word-formation of Distributed Morphology (DM). The resulting framework, which we call Lexical-Realizational Functional Grammar (LRFG), is particularly well-suited to modelling Canadian Indigenous languages, which are characterized by polysynthesis and nonconfigurationality. In this paper we summarize the framework, and demonstrate it with an analysis of the inflectional system of Ojibwe, a language showing these properties. Note that the intent of the paper is not to make new claims about Ojibwe, but instead to take existing descriptions of the language and adapt them to the present formalism as a demonstration of the framework.