In the English lexicon, laryngeal alternations in the plural (e.g. leaf ~ leaves) impact monosyllables more than finally stressed polysyllables. This is the opposite of what happens typologically, and would thereby run contrary to the predictions of initial-syllable faithfulness. Despite the lexical pattern, in a wug test we found monosyllables to be impacted no more than finally stressed polysyllables were—a ‘surfeit of the stimulus’ effect, in which speakers fail to learn a statistical generalization present in the lexicon. We present two artificial-grammar experiments in which English speakers indeed manifest a universal bias for protecting monosyllables, and initial syllables more generally. The conclusion, therefore, is that speakers can exhibit spontaneous learning that goes directly against the evidence offered by the ambient language, a result we attribute to formal and substantive biases in phonological acquisition.

This article proposes a new way of understanding grammatical status and grammaticalization as distinctive types of linguistic phenomena. The approach is usage-based and links up structural and functional, as well as synchronic and diachronic, aspects of the issue. The proposal brings a range of previously disparate phenomena into a motivated relationship, while certain well-entrenched criteria (such as ‘closed paradigms‘) are shown to be incidental to grammatical status and grammaticalization. The central idea is that grammar is constituted by expressions that by linguistic convention are ancillary and as such discursively secondary in relation to other linguistic expressions, and that grammaticalization is the kind of change that gives rise to such expressions.

We propose a new approach to metrics based on maxent grammars, which employ weighted constraints and assign well-formedness values to verse lines. Our approach provides an account of metricality and complexity that has a principled mathematical basis and integrates information from all aspects of scansion. Our approach also makes it possible to detect vacuous constraints through statistical evaluation.

We begin with a system built on earlier work that defines the set of possible constraints, following principles of stress matching, bracket matching, and contextual salience. The basic concepts of this system work well in describing our data corpora, taken from Shakespeare’s Sonnets and Milton’s Paradise Lost. However, one well-known type of constraint, based on the principle of the stress maximum (Halle & Keyser 1966 et seq.), emerges as vacuous; testing indicates that the work of such constraints is already done by simpler constraints independently needed in the grammar.

Deaf children whose hearing losses are so severe that they cannot acquire spoken language and whose hearing parents have not exposed them to sign language nevertheless use gestures, called HOMESIGNS to communicate. Homesigners have been shown to refer to entities by pointing at that entity (a demonstrative, that). They also use iconic gestures and category points that refer, not to a particular entity, but to its class (a noun, bird). We used longitudinal data from a homesigner called David to test the hypothesis that these different types of gestures are combined to form larger, multigesture nominal constituents (that bird). We verified this hypothesis by showing that David’s multigesture combinations served the same semantic and syntactic functions as demonstrative gestures or noun gestures used on their own. In other words, the larger unit substituted for the smaller units and, in this way, functioned as a nominal constituent. Children are thus able to refer to entities using multigesture units that contain both nouns and demonstratives, even when they do not have a conventional language to provide a model for this type of hierarchical constituent structure.

The case-number suffixes of the Western Nilotic language Nuer (Frank 1999) display a remarkable combination of formal simplicity and distributional complexity, which is manifested in: (i) a seemingly erratic form-function mapping that precludes attributing a consistent meaning to the suffixes, and (ii) a wealth of inflection classes only barely differentiated from each other. The suffixes looks as if they were rule-generated, but behave as if they were memorized. I advance a model of inflection combining principal parts, implicational rules, and default inheritance, in which the bulk of the complexity is attributed to the lexical stem, revealing the underlying systematicity behind suffix assignment.

Children (aged five-to-six and nine-to-ten years) and adults rated the acceptability of well-formed sentences and argument-structure overgeneralization errors involving the prepositional-object and double-object dative constructions (e.g. Marge pulled the box to Homer/Marge pulled Homer the box). In support of the entrenchment hypothesis, a negative correlation was observed between verb frequency and the acceptability of errors, across all age groups. Adults additionally displayed sensitivity to narrow-range semantic constraints on the alternation, rejecting double-object dative uses of novel verbs consistent with prepositional-dative-only classes and vice versa. Adults also provided evidence for the psychological validity of a proposed morphophonological constraint prohibiting Latinate verbs from appearing in the double-object dative. These findings are interpreted in the light of a recent account of argument-structure acquisition, under which children retreat from error by incrementally learning the semantic, phonological, and pragmatic properties associated with particular verbs and particular construction slots.

A variety of mechanisms have been proposed in sociolinguistics for the propagation of an innovation through the speech community. The complexity of social systems makes it difficult to evaluate the different mechanisms empirically. We use the four-way typology of mechanisms proposed by Baxter and colleagues (2009), and define them mathematically in such a way that the effects of different mechanisms in the trajectory of a change can be modeled. The model suggests that the widely observed empirical pattern of an S-curve temporal trajectory of change can be captured only if the mechanisms for propagation include replicator selection, that is, differential weighting of the competing variants in a change, except under highly specialized circumstances that probably do not hold in speech communities in general.

Same-except is a fundamental, domain-general cognitive relation in which entities in proximity to one another are judged to be the Same, Except for some part or property where they differ. This relation can be attested in nonlinguistic modalities such as vision, audition, and taste, and it plays an important role in nonlinguistic categorization. We show that this relation is expressed linguistically by means of a wide range of devices, including (a) lexical expressions such as same and except, (b) contrastive stress, (c) anaphora (e.g. definite and indefinite NP anaphora and VP anaphora), (d) ellipsis (e.g. bare argument ellipsis, sluicing, gapping, and VP ellipsis), and (e) fixed expressions such as vice versa and that goes for. Our approach thereby unifies the semantics of all of these phenomena under a common account that is based on a domain-general cognitive principle. We compare our approach with accounts of ellipsis based on syntactic copying or deletion, showing that although both approaches have their difficulties, the challenges to the copying/deletion approach are more severe.

This article reexamines a controversial construction in Acehnese (Lawler 1977 versus Durie 1988). I demonstrate that the construction is a passive, even though a verbal prefix bears the features of the agent rather than the surface subject. I analyze the prefix as a morphological realization of the functional head that introduces the external argument; the features borne by this head are not agreement, but rather interpretable features that restrict the external argument position. Important consequences are that Acehnese does not counterexemplify the universality of grammatical relations (contra Durie 1988 and subsequent), and that Acehnese provides clear morphological evidence for the presence in passives of the functional head that introduces the external argument.

The source of syntactic island effects has been a topic of considerable debate within linguistics and psycholinguistics. Explanations fall into three basic categories: grammatical theories, which posit specific grammatical constraints that exclude extraction from islands; grounded theories, which posit grammaticized constraints that have arisen to adapt to constraints on learning or parsing; and reductionist theories, which analyze island effects as emergent consequences of non-grammatical constraints on the sentence parser, such as limited processing resources. In this article we present two studies designed to test a fundamental prediction of one of the most prominent reductionist theories: that the strength of island effects should vary across speakers as a function of individual differences in processing resources. We tested over three hundred native speakers of English on four different island-effect types (whether, complex NP, subject, and adjunct islands) using two different acceptability rating tasks (seven-point scale and magnitude estimation) and two different measures of working-memory capacity (serial recall and n-back). We find no evidence of a relationship between working-memory capacity and island effects using a variety of statistical analysis techniques, including resampling simulations. These results suggest that island effects are more likely to be due to grammatical constraints or grounded grammaticized constraints than to limited processing resources.

In this article, we consider the binding-theoretic status of the Korean long-distance anaphor caki. While caki has been called a long-distance anaphor, in reality its antecedent can be found locally as well as at a distance. It can also have a non-c-commanding antecedent, and an antecedent from a previous sentence. Though there are many different approaches to caki, what is apparent is the generalization that caki must be coindexed with an NP/DP if there is a possible antecedent in the syntax. We take this one step further and show that caki must be bound if there is a possible binder in the semantics, using examples where caki is bound by implicit arguments coming from reportative evidential and generics/modals. We argue that this generalization is best captured if caki is seen as a bound variable requiring a semantic binder, and demonstrate how this bound-variable analysis can provide a unified account for local, long-distance, and discourse-bound instances of caki, as well as instances of caki with a non-c-commanding antecedent and those bound by an implicit argument. The residual cases where caki has no possible semantic binder are treated as instances of exempt anaphora, free variables, the felicity of which are subject to discourse conditions.

That there is a theoretical distinction between context-dependent and context-independent aspects of utterance interpretation has become a standard assumption in current theories of meaning; however, how and where to draw this distinction has been the subject of considerable debate. In the current study, we investigate whether speakers can systematically distinguish between what is said and what is implicated (Grice 1989 [1967]) using a novel truth-value judgment paradigm across a wide range of implicature types. We found that, by providing participants with a clear set of judgment criteria, including the adoption of an objective third-person perspective, we were able to enhance their ability to distinguish conversational implicature from truth-conditional meaning. In addition, we found that none of the implicature types we investigated was either consistently incorporated into or consistently excluded from what is said. Instead, our findings revealed considerable variation in frequency of incorporation across implicature types in ways that do not correspond straightforwardly to the various taxonomies of implicature types proposed in the literature.

In this article, I focus on the intonation patterns of Turkish-German bilinguals to discuss intonation within the context of language contact and language variation. The intonational variance involves the realization of terminal rises as produced by second- and third-generation Turkish-German bilinguals living in Germany. These speakers produce two phonetically, phonologically, and pragmatically distinct rises, which differ from what is typically reported for German monolinguals. The primary phonetic differences between the two rises include the relative alignment and slope of the rise, with one rise aligning on the final syllable of the word regardless of the stress pattern and showing a significantly steeper slope than the other. Although the source of these two rises is likely the two languages used by the speakers, this is not a case of intonational code-switching. Rather, the two rises, along with other edge phenomena, form an intonational system in which the rises are in contrast with one another as well as with falls and level edge contours and as such play different pragmatic roles relative to one another.

Much recent work argues that lexical frequency plays a central explanatory role in linguistic theory, but the status, predicted effects, and methodological treatment of frequency are controversial, especially so in the less-investigated area of syntactic variation. This article addresses these issues in a case study of lexical frequency effects on variable subject personal pronoun (SPP) expression in Spanish. Prior studies of Spanish SPP use revealed significant constraints including formal and semantic properties of the verb, and discourse factors such as a switch reference. These constraints appear to be confirmed in our analysis of 4,916 verbs from a spoken corpus of Spanish, as is a powerful role for lexical frequency. But the frequency effect—best configured as a discrete rather than continuous variable—is complex; statistically, it has no independent direct effect, but operates entirely through interaction with other constraints. All other constraints on SPP use are amplified in high-frequency forms, and some disappear at low frequencies. Frequency thus acts as a ‘gatekeeper’ and potentiator: above some frequency threshold, significant linguistic constraints on SPP use emerge; below the threshold they do not. We propose that this reflects experience and acquisition: speakers cannot formulate hypotheses about individual lexical items until they have sufficient evidence; the threshold is the level at which speakers have enough experience with a form to do so.

These results have important theoretical and methodological implications. They require rich lexical representations incorporating frequency and collocational information. Methodologically they indicate the need for careful quantitative explorations of frequency, because its role as an enabler of other constraints produces unstable statistical results.

This article compares an extensive collection of English loanwords into Korean with a corpus of perceptual responses to English productions by Korean students of English in Korea. The analysis selects ten obstruents situated in four prosodic contexts: initial, final, and pre- and poststress intervocalic positions. Analyses compare the mapping of the obstruents onto Korean categories in the two databases, finding a strong logistic relationship between them, which indicates a process of loanword adaptation as a regularization of the cross-language perception patterns. This conclusion is also supported by differences in the maps across the prosodic positions, wherein loanword differences are correlated with perceptual differences, and by the fact that loanword adaptations are more variable for consonants that do not have a very robust perceptual map. The data, however, also exhibit exceptional cases that apparently indicate effects of a historical lexicalization of individual forms, as well as of an explicit sociocultural standard. Thus, loanword adaptation in this case, though largely indicative of a perceptual base, is more than just synchronic perception.

The goal of this discussion note is to argue that the complex facts of hybrid agreement in Serbo-Croatian can and should be accounted for by assuming that only two sets of features are relevant to agreement: a syntactic set and a semantic set. This is in opposition to existing proposals that claim that two distinct sets of syntactic features are required in addition to the set of semantic features in order to account for those agreement facts (Wechsler & Zlatić 2000, 2003). The proposal defended here with two sets of agreement features is shown to be superior to the alternative with three feature sets because it not only is simpler and accounts for the facts just as well, but it also does not make some incorrect predictions that the alternative makes when crucial facts are considered.

We present the first proposal of detailed internal subgrouping and higher-order structure of the Pama-Nyungan family of Australian languages. Previous work has identified more than twenty-five primary subgroups in the family, with little indication of how these groups might fit together. Some work has assumed that reconstruction of higher nodes in the tree was impossible, either because extensive internal borrowing has obscured more remote relations, or because the languages are not sufficiently well attested (see, for example, Bowern & Koch 2004b, Dixon 1997). With regard to the first objection, work by Alpher and Nash (1999) and Bowern and colleagues (2011) shows that loan levels are not high enough to obscure vertical transmission for all but a few languages. New data remove the second objection. Here we use Bayesian phylogenetic inference to show that the Pama-Nyungan tree has a discernible internal subgrouping. We identify four major divisions within the family and discuss the implications of this grouping for future work on the family.

While all of the Scandinavian languages have verb-second order in main clauses, they vary in the word order in subordinate clauses: in Icelandic the finite verb appears in a high position, to the left of negation and sentence-medial adverbs, while in all of the standard Mainland Scandinavian languages it remains in a low position, to the right of these elements. This order in Mainland Scandinavian is known to be the result of a historical change, and has frequently been tied to the loss of agreement morphology. Faroese has been argued to be currently undergoing a change of the same type, but it has proved difficult to establish a sound empirical footing for the various claims about the syntax of this language. In this article we present data from three experimental investigations of acceptability, supplemented with a study of available texts, that show that the language is very close to completing the change in the loss of the high position for the verb, but that its syntax is still distinct in this respect from that of Danish, the mainland Scandinavian language with which it is in most contact. In addition to establishing a firmer empirical basis for theories of verb movement, our study also makes the methodological point that grammaticality-judgment tasks can yield extremely fine-grained results even in cases where variability is at issue.

Linguists have increased their documentation efforts in response to the sharp decline in the number of languages. Greater awareness and new sources of funding have led to an upsurge in language documentation. While individual languages make unique contributions to the world's linguistic heritage, language families, by virtue of their shared heritage, have related contributions to make. The endangerment of entire families, while implied by the loss of language, has not been explored to date. Here, we examine estimates of how many linguistic stocks (the largest subgroups of related languages that are reconstructable) consist entirely of endangered languages and thus are endangered themselves. Our data set includes 372 stocks that had at least one living language in 1950. Our finding is that since that time, 15% of the world's linguistic stocks have become extinct and another 27% are now moribund in that direct estimates of endangerment indicate that no member languages are being learned by children. For cases where direct estimates were not found, we used population as a proxy for endangerment. If, as many predict, a further 50% of the world's languages become extinct or moribund in this century, then an additional 25% of the linguistic stocks would be at risk of being lost. If 90% share this fate, only 11% of stocks would have at least one presumably ‘safe’ language. A comparison of the vitality of linguistic stocks by world areas yields the following ranking from worst to best: Americas, Pacific, Asia, Africa, Europe. Finally, some aspects of language that are unique to disappearing language stocks are outlined. Renewed efforts at documenting members of such stocks seem justified.