Chapter 8 discusses alternative theories of case, in particular the few prior attempts to deal with the Faroese dative–accusative pattern (Woolford 2007, Jónsson 2009, Asarina 2011). Further survey data from Faroese are presented alongside engagement with these alternatives; it is argued that while these theories could be altered to achieve empirical coverage, they will miss generalisations and overgenerate in comparison to the OLG approach. Woolford (2007) deals with exceptions to Burzio’s Generalisation (Burzio 1986); where Woolford’s account runs into problems is the conflation of abstract and morphosyntactic case: the case-hierarchy constraints alone do not explain the possibility of mismatches between thematic structure, argument structure and case-marking on arguments. Second, an idea proposed by Jónsson (2009), built upon by Asarina (2011), is discussed:‘covert’ nominative case. The basic idea is not dissimilar to abstract [+HR] case instantiated in morphosyntax as the case borne by subject position; however, crucial differences render the covert nominative account undesirably over-flexible. Further survey data are presented with respect to purported nominative ‘objects’, showing that such sentences are unacceptable in contemporary Faroese and that the data are inconsistent with Asarina (2011).

Chapter 9, the longest chapter, presents a step-by-step discussion of the OLG theory of syntax, focusing especially on issues of central concern to syntacticians: phrase structure, movement or filler-gap dependencies, and the architecture of grammar. A detailed walk-through of how to derive an English sentence is given, including formal definitions of syntactic features. Analyses of the range of typological variation observed in relevant word-order, case-marking and information-structural phenomena are presented from an OLG perspective, including detailed case studies of the Faroese clause structure facts presented in Chapter 2 and an in-depth treatment of object shift in Scandinavian languages. Ranking arguments, constraint definitions and factorial typologies are given where needed. Chapter 9 is intended to answer most of the major questions regarding how this theory handles a broader range of data.

This chapter focuses on an illustrative phenomenon that has presented challenges for previous theories: non-nominative subjects. A summary of preceding literature is given, highlighting the relevant subjecthood properties of the Icelandic preverbal datives, as well as the subject-licensing syntactic positions in that language. A contrast is then drawn with similar dative arguments in German, concluding that the German obliques do not behave as subjects with respect to control properties, and do not occupy Spec,TP like the Icelandic datives. Before investigating the subjecthood of the Faroese datives, a detailed overview of what is currently known of Faroese clause structure is presented to establish the evidence for argument-licensing positions in the language. The standard subjecthood tests are applied to the Faroese dative-experiencer predicates, which demonstrate that the Faroese datives behave very similarly to those in Icelandic with respect to subjecthood properties. Given that the dative arguments in such sentences appear to be true subjects, and therefore the subject licensing position of Spec,TP seems to be the same in both Faroese and Icelandic, the differences in object case and number agreement with a plural object remain to be explained. This is explored in Chapters 3–7.

Chapter 7 lays out new data from both Faroese and Icelandic regarding triadic verbs, in particular the passive of ‘give’ and other three-place predicates. The theoretical apparatus presented in Chapter 6 is brought to bear on ditransitives and shown to predict the correct case frames and word orders in Icelandic and Faroese. First, an outline is given of double-object verbs in Faroese, noting that the evidence in some of the preceding literature is equivocal as to the acceptability of passive with ‘give’ and other triadic verbs. Data from a Faroese survey are discussed, the result being that no sentence with passive of ‘give’ was broadly accepted. Faroese evidence is discussed regarding the position of the theme and goal arguments in the active. Further data on the ‘give’ passive in Icelandic are presented; consistently with previous work, these Icelandic speakers have the option of either Goal-V-Theme or Theme-V-Goal orders in the passive. A Faroese survey on passives of ditransitives other than ‘give’ shows that the lexical semantics of a given verb interact with word order, such that if passive is judged acceptable, its mean acceptability is higher for the Theme-V-Goal order than for Goal-V-Theme.

In Chapter 5, the competing grammars model of morphosyntactic variation is introduced from both a sociolinguistic and computational perspective. The example of nominative substitution in Faroese is used to demonstrate the advantages of the model, in particular the combination of classic Optimality Theory constraints with a probabilistic activation hypothesis. The Faroese dative-subject verbs discussed in Chapters 2 and 4 occur in both dative–accusative and nominative–accusative case frames. The competing grammars model is outlined as a cogent explanation of the co-existence of both forms in use by a given speaker, sometimes within the same text or short series of utterances. Relevant factors proposed to influence selection of the nominative versus dative variants are discussed, including both grammatical and social/contextual variables. The importance of social meaning in determining case selection is highlighted, which presents a Rational Speech Act model of this morphosyntactic variable. In a section co-authored with Rob Mina, the issue of bimodally distributed judgement data is explored, in particular whether such data are effectively random or represent distinguishable dialects, and how to tell. Finally, neural approaches are discussed as an alternative model of competing grammars.

The introduction presents a general discussion of syntactic theories, contrasting transformational approaches with those that adopt feature-structural representations to show how they provide differing cognitive models of the human language faculty. The three key components of the OLG framework – Linking Theory (LT), Optimality Theory (OT) and Competing Grammars (CG) – are briefly laid out. These are motivated by phenomena that necessitate multiple levels of case, linking between levels that permits mismatches in some grammars, a means of restricting the linking apparatus while still capturing the data, and a cogent account of morphosyntactic variation. One such phenomenon is introduced: case-marking facts in Insular Scandinavian. The dative–nominative Icelandic predicates are contrasted with the Faroese dative–accusative pattern, along with plural number agreement with the object in Icelandic versus non-agreement in Faroese. Next, an overview is presented of the motivation for each theoretical component of OLG, in turn outlining the advantages of LT, OT and the CG hypothesis. The introduction concludes with an outline of the specific empirical findings from surveys conducted on the Faroe Islands and Iceland, including quirky case predicates and passives, followed by an overview of the book’s structure.

Chapter 4 zooms in on dative-subject predicates in Faroese and presents new survey data from the Faroe Islands and Iceland. The preceding literature on Faroese non-nominative subjects is reviewed before the Faroese and Icelandic surveys on quirky case monotransitives are described in detail. The implications of the results are discussed in relation to case-assignment, agreement and word order, and the author’s proposed analysis is presented along with a factorial typology. Two Faroese surveys and one Icelandic survey testing possible object positions in quirky case sentences are discussed. The results are consistent with the hypothesis that the structural object position of nominative arguments in Icelandic, and that of accusatives in Faroese, is the same regular object position of nominative–accusative case frames. On the other hand, the results are inconsistent with the hypothesis that the object position is different between the two languages (and hence that such a difference is responsible for the difference in case-marking). These conclusions are discussed in detail in the OLG analysis section, which attributes the difference to constraint interaction, particularly a different ranking of a pair of constraints enforcing structural object case (accusative) and agreement with a nominative argument, respectively.

Chapter 3 outlines OLG’s core components, giving the prerequisite theoretical background for understanding the data analyses that follow. The Linking Theory section answers questions regarding levels of lexical–semantic and syntactic representation, mapping between levels, and the architecture of grammar. A short review is presented of the core propositions of Optimality Theory (OT) approaches to syntax, including a list of proposed constraints governing case-assignment. Linking Theory originates in Kiparsky (1997); the central innovation is that three levels of case – abstract, morphosyntactic and morphological – are distinguished, and the same set of binary features is seen to operate at all levels, though with level-appropriate realisation. The introduction to OT focuses on its application to the syntactic module of grammar; a key point is that a huge number of unlikely candidates are harmonically bounded by undominated or high-ranked constraints. Such markedness constraints interact with faithfulness constraints enforcing realisation of all features present in the input, preventing omission of input material other than in highly marked forms. This allows us to capture long-standing generalisations about phrase structure without having to claim they will be completely unviolated across languages: language-specific rankings may result in more or less marked structures.

Chapter 6 presents new data on Faroese personal and impersonal passives, in addition to a discussion of case preservation and availability of passive with the dative-subject verbs. First, data from further Faroese surveys are investigated. The analysis demonstrates that the proposed constraints regulating the passive, here simply the addition of PARSE, covers all relevant sentence types in both Faroese and Icelandic personal passives, and with DEP and ARGSP also the impersonal passives. The argumentation builds on Kiparsky (2013), providing additional empirical support for the Linking Theory approach. An important finding is that the facts of case substitution in the active and case non-preservation in the passive are related: it is argued that there are preserving versus non-preserving grammars, represented by the two competing rankings. Speakers have access to both rankings, and choice of verb strongly predicts which grammar is activated. This approach predicts patterns that emerge from the constraint rankings themselves. The non-preserving grammar (activated, for example, with uses of the verb ‘like’) implies availability of passive and nominative substitution, both of which hold true; likewise, the preserving grammar predicts unavailability of passive and case substitution, which again turn out to be true for the verb ‘need’.

Within a system, multiple patterns of rule combination may interact in complex ways. I present a detailed analysis of Swahili verb inflection in which simple rules, composite rules and aggregated rules all enter into intricate competition, yielding an extravaganza of deviations from canonical morphotactic criteria. At the center of this discussion are three characteristics of Swahili conjugation: (i) the polyfunctionality of verbal concords (in virtue of which the same rule is used to express the noun class of a verb’s subject, that of its pronominal object, or that of a relativized argument), (ii) the expression of negation (by means of three complementary rules), and (iii) the marking of relative verb forms (whose relativized-argument affix participates in an extensive pattern of affix counterposition). The rule-combining approach to morphotactics allows the interacting details of these subsystems to be resolved into two very general types exhibiting an unexpected degree of economy.

I discuss the general implications of the rule-combining approach to morphotactics developed in the course of the foregoing chapters. I summarize the numerous superficially problematic phenomena that the rule-combining approach resolves and I relate these phenomena to the variety of ways in which rule combinations may deviate from the canonical characteristics of a language’s morphotactics. Finally, I synopsize the set of formal definitions on which the rule-combining approach is based.

Besides affording a way of modeling deviations from canonical morphotactics, rule composition makes it possible to see apparently recalcitrant morphotactic patterns as conforming to canonical criteria if these are assumed to cover composite rules as well as simple rules. I examine an apparent deviation from the integral stem criterion in Sanskrit and apparent deviations from the rule opposition criterion in Latin, Limbu, and Sanskrit. Each of these phenomena can be reconciled with the canonical criterion from which it apparently deviates if this criterion is assumed to cover composite rules as well as simple rules. All of these are cases in which deviation from the minimal rule criterion facilitates conformity to other canonical morphotactic criteria.