The Form of the Phi Kernel

This chapter considers the broader lessons that might be drawn from the current study of person. It draws out the core commonalities between the person and number features that have been proposed: that features are “operations” richer than first-order predicates held together by conjunction; that they are not subject to extrinsic constraints on order of composition or co-occurrence; and that there are semantic and morphological grounds for representing features of both kinds bivalently. The consequences of this study might ramify beyond linguistics by altering our understanding of and means of investigating the language of thought and the nature and evolution of mind. The author asserts, contrary to widespread opinion, minds do leave fossils, but these are to be sought, not by paleoanthropologists sifting through the archeological record, but by cognitive scientists, including linguists, via our theories of the structure of the mind itself.

Oldfangled and

Traditional accounts of person assume that features denote first order predicates, that their values denote one-place truth functors, and that feature bundles are held together semantically by conjunction. Crucially, conjunction is a commutative operation, unlike those belonging to the current theory. The current chapter explores the consequences of semantics commutativity for theories of person features. Reviewing a range of influential accounts, it shows that these are accounts undergenerate if given only two features, but overgenerate if given more, and that means of trimming the generative excess are unsatisfactory. The chapter also compares three analyses of Bininj-Gunwok, which has a tripartite person for objects but quadripartite for person, arguing that the smallest feature inventory yields the most compact account.

Morphological Composition

Arguments from morphological compositionality in support of feature analyses are common in treatments of number. This chapter argues that they have an important contribution to make to person and adduces a partial typology that strongly supports the analysis of the previous chapter. Particularly noteworthy patterns are inclusives that subsume exclusives, exclusives that subsume inclusives, first persons that subsume second, second persons that subsume first, and second persons that arise from crosscutting first and third.

The Partition Problem

With the previous chapter having established that partitions are the correct object of fundamental inquiry for the theory of person, the current chapter presents the partition chapter in its most general form: essentially, why only five of fifteen (or more) possible partitions are attested. This empirical basis of the problem involves person partitions of all sizes as well as person-related partitions also of spatial deixis (spaces, objects, paths).

The Path to Partition

Morphological theory generally takes syncretism to be a key source of evidence about features and natural classes. The current chapter argues against this status quo, showing that there are no categorical constraints on possible versus impossible syncretisms. Partitions, by contrast, offer a means of marshalling the data productively. Defined as the total set of distinctions made by all paradigms within a given grammatical domain, partitions transform person in a categorical problem of the attested versus the unattested. Focusing narrowly on three-person pronominal systems, the chapter shows that the skewed distribution of partitions is not attributable to functional or historical factors and demands a deeper explanation.

Spaces, Objects, Paths

Generative work in morphology does not normally pursue a connection between person and spatial deixis, unlike more functionalist and cognitive linguistic work. This chapter presents three kinds of evidence for this move, however: morphological parity between partitions, semantic parity between partitions, and the crosscutting of person- and non-person-based spatial distinctions. A sketch of the syntax and semantics of the personal basis of spatial deixis is also presented.

Conclusion

This investigation began with the very modest question of why, in languages like English, the meaning ‘we and you’ is covered by we, not you. By broadening this question to all partitions of person and related deictic spaces, we realize that there is a substantial disparity between possible and attested systems of person and person-related deixis. Such shortfall between the possible and the actual is a classic variety of linguistic problem, and, in responding to it, I have attempted to present a very minimal theory of person. The resulting system furnishes a range of desirable consequences, concerning the morphological composition of different persons, the semantic and morphological interaction of person with number, and the capacity to capture the relationship between personal and spatial deixis. What emerges is a coherent, unified view of the kernel of phi, the features that make up person and number. These results may ramify beyond phi theory into the linguistic theory of other feature families and beyond, into broader issues in cognitive science and the evolution of mind. The study of phi features brings us to a deeper understanding, not just of what we is as a pronoun, but of who we are as thinking creatures.

The Partition Problem Solved

This chapter solves the partition problem of the previous chapter. It posits an ontology which is organised into nested power sets, denoted by two features (author, participant) and their host head. Crucially, these features are bivalent and their values denote operations between sets, allowing one set-denoting feature to act on another. The chapter shows that, once the systems can account for the most basic partitions, all and only the others follow, depending on a semantic parameter of order of composition.

Number and the Functional Sequence

A very basic demand on any theory of person is that it interact with a theory of number to yield familiar inventories of pronouns and agreement in a straightforward fashion. This chapter shows that the current theory does so. Specifically, it shows that the theory of Harbour 2014 (Paucity, abundance, and the theory of number, Language) generates singular, dual, trial, paucal, greater paucal, minimal, augmented, unit augmented, and plural, both with and without clusivity contrasts. The chapter also argues for a phrase structure in which number dominates person and applies novel data from person to the question of whether the denotation of plural includes or excludes that of singular.

In a Nutshell

This chapter presents the core empirical problem in terms of English pronouns and lays out the three central theses of the book. The empirical thesis is that the person problem is to be formulated in terms of partitions, not syncretisms, and must attend to person and spatial deixis, not to person (as traditionally defined) alone. The theoretical thesis is that there are just two person features, that they and their syntactic host denote power sets of subsets of the ontology, and that their values denote operations between lattices (sets). The metatheoretical thesis generalizes the understanding of features beyond person.