Annual Review of Linguistics - Volume 3, 2017

William Labov launched the branch of language studies known as language variation and change or (more vaguely) sociolinguistics with several influential studies in the 1960s and seminal publications in the early 1970s. As the discipline spread to universities and research institutes in North America and, with remarkable rapidity, to Europe, Asia, the Antipodes, and ultimately all corners of the globe, Professor Labov has remained at or near the forefront of its various movements, including historical sociolinguistics, narrative analysis, sociometrics, subjective evaluation testing, educational and forensic applications, and sociophonetics, as well as mainstream variation studies. His breadth and depth as a scholar have influenced countless linguists and policy makers, and his enthusiasm and boundless optimism have inspired numerous colleagues, collaborators, and students. Several of them relate their personal experiences with the man and his ideas here.

This article reviews the role phonology plays in phonetic changes. After first establishing what kinds of changes qualify as phonetic changes for the purposes of discussion, and laying out the theoretical outlook that is adopted here, I review the most obvious cases in which phonology plays a role in phonetic change. These include (a) the way phonological contrast can lead to phonetic dispersion, (b) the way phonological natural classes can define a set of segments to undergo a parallel phonetic shift, and (c) how phonological biases may lead to instances of underphonologization. Throughout, I discuss alternative approaches to these phenomena.

Comparing phonology in spoken language and sign language reveals that core properties, such as features, feature categories, the syllable, and constraints on form, exist in both naturally occurring language modalities. But apparent ubiquity can be deceptive. The features themselves are quintessentially different, and key properties, such as linearity and arbitrariness, although universal, occur in inverse proportions to their counterparts, simultaneity and iconicity, in the two modalities. Phonology does not appear full blown in a new sign language, but it does gradually emerge, accruing linguistic structure over time. Sign languages suggest that the phonological component of the language faculty is a product of the ways in which the physical system, cognitive structure, and language use among people interact over time.

The multifacetedness of morphological complexity stems partly from the fact that linguistic complexity can be construed in more than one way, but also from the very nature of morphology. The complexity of morphological systems can be compared with respect to their morphotactics, to their word forms’ morphosyntactic content, to their exponence relations, to their derivational semantics, to their reliance on morphomic patterns, to the implicative relations among the forms in their inflectional paradigm, to their interfaces with other grammatical components, and to the relative parsability of their word forms. Recent research on morphological complexity has led to a precise proposal for quantifying complexity in some of these different dimensions; some dimensions, however, lend themselves to quantification more straightforwardly than others.

This article reviews research on the unsupervised learning of morphology, that is, the induction of morphological knowledge with no prior knowledge of the language beyond the training texts. This is an area of considerable activity over the period from the mid 1990s to the present. It is of particular interest to linguists because it provides a good example of a domain in which complex structures must be induced by the language learner, and successes in this area have all relied on quantitative models that in various ways focus on model complexity and on goodness of fit to the data.

Learnability has traditionally been considered to be a crucial constraint on theoretical syntax; however, the issues involved have been poorly understood, partly as a result of the lack of simple learning algorithms for various types of formal grammars. Here I discuss the computational issues involved in learning hierarchically structured grammars from strings of symbols alone. The methods involved are based on an abstract notion of the derivational context of a syntactic category, which in the most elementary case of context-free grammars leads to learning algorithms based on a form of traditional distributional analysis. Crucially, these techniques can be extended to work with mildly context-sensitive grammars (and beyond), thus leading to learning methods that can in principle learn classes of grammars that are powerful enough to represent all natural languages. These learning methods require that the syntactic categories of the grammars be visible in a certain technical sense: They must be well characterized either by the sets of symbols that they generate or by the sets of contexts in which they can appear. However, there are still significant gaps between these theoretical results and their direct implementation as models of language acquisition; I discuss how these remaining problems can be overcome.

This review introduces and compares syntactic structures in a variety of sign languages. I first examine ways in which sign languages function like spoken languages, and ways in which they differ. I then briefly discuss what sign languages have in common in the syntactic realm; the rest of the article focuses on how they can differ. Because the level of the simple sentence has been documented extensively, this review emphasizes complex sentences, such as sentential complementation, relative clauses, adverbial clauses, embedded questions, and conditionals.

Sign language anaphora is often realized very differently from its spoken language counterpart. In simple cases, an antecedent is associated with a position or “locus” in signing space, and an anaphoric link is obtained by pointing toward that locus to recover its semantic value. This mechanism may sometimes be an overt realization of coindexation in formal syntax and semantics. I discuss two kinds of insights that sign language research can bring to the foundations of anaphora. First, in some cases the overt nature of indices in sign language allows one to bring overt evidence to bear on classic debates in semantics. I consider two: the availability of situation-denoting variables in natural language and the availability of binding without c-command. Second, in some cases sign language pronouns raise new challenges for formal semantics. Loci may function simultaneously as formal variables and as simplified depictions of what they denote, requiring the construction of a formal semantics with iconicity to analyze their properties.

The fundamental importance of lexical categories is uncontroversial within both formal and functional approaches to grammatical analysis. But despite the familiarity of this topic and its foundational nature for grammatical description and analysis, it is paradoxically not among the best-studied or -understood topics from either the functionalist or formalist perspective. Both schools of linguistic theory have inherited their basic assumptions and instincts about lexical categories from the structuralist practice of distributional analysis. We briefly survey approaches to the various lexical categories. We then comment on a few issues of strategic value that arise from these approaches, including the importance of clearly distinguishing roots, stems, words, and syntactic units when it comes to issues of lexical categories; the importance of recognizing when distributional tests are similar across languages in principled ways; and the need for the choice of distributional tests to be informed by theoretical hypotheses.

The count/mass distinction is a domain of human language that is strongly related to human cognition. This review starts with an overview of recent research on individuation in relation to core knowledge systems (in particular the systems of object representation, agent representation, and number representation), followed by a brief overview of the role these representations play in the acquisition of count meaning. I then discuss linguistic aspects of individuation and the count/mass distinction in more detail. I distinguish between the grammatical systems of languages and the lexical properties of nouns, focusing on crosslinguistic variation. Languages vary substantially in the grammatical conditions that quantity expressions impose on the nouns they combine with, as well as in the exact lexical and grammatical properties of nouns. At the same time, individuation and counting seem to play a role in all languages, and similar counting strategies show up in unrelated languages.

Explaining children's nonadult interpretations of sentences with quantifiers has been the objective of extensive research for more than 50 years. This article reviews four areas of research, each of which began with the observation that children and adults respond differently to sentences with quantifiers. The observed differences have been subject to considerable debate, often drawing upon linguistic theory for answers and sometimes resulting in changes to the theory. This article begins by discussing children's comprehension of sentences with pronouns with quantificational versus referential antecedents. The next topic is children's nonadult responses to sentences with quantifiers and negation. The third topic is children's analysis of scope phenomena. I conclude with a discussion of children's understanding of the focus adverb only, which is used to expose some common properties of historically distinct languages. Progress in each of these four areas has revealed children's deep understanding of the basic meanings of quantifiers and how quantifiers interact with other logical expressions. I conclude that children's nonadult interpretations of quantifiers are consistent with the theory of Universal Grammar.

Scalar implicatures (SIs) and, more generally, quantity-based implicatures (QBIs) have been intensely investigated since Grice's seminal work. Recently, SIs and QBIs have been at the center of an intense debate. Some researchers, following Grice's original insight, argue that they should be captured solely in terms of principles of rational action (the pragmatic approach). Others argue that they cannot be analyzed in purely pragmatic terms but can only be properly understood in terms of a compositional semantic device, namely exhaustification (the grammatical approach). In this article, I review the key arguments in this debate, which is of interest not only to determine who is right but also because of the range of new phenomena that have come to light thanks to such a debate. My conclusion is that both conceptual and empirical reasons favor the grammatical approach.

We discuss problems familiar from the literature on presupposition and information structure, and illustrate how a synthesis using the Question Under Discussion (QUD) framework yields fresh insight. In this framework, discourse is analyzed in terms of the strategy of inquiry pursued by the interlocutors, and individual utterances are interpreted relative to the question being addressed. This way of thinking offers a new perspective on diverse phenomena, including the projection of presuppositions, association with focus, contrastive topic marking, and variability of projection behavior. We review the principal issues and prior lines of research in each of these areas, and show how the issues may be recast in QUD terms.

Children from low-income backgrounds consistently perform below their more advantaged peers on standardized measures of language ability, setting long-term trajectories that translate into gaps in academic achievement. Our primary goals in this review are to describe how and why this is so, in order to focus attention on ways to enrich early language experiences across socioeconomic strata. We first review the literature on the relation between socioeconomic status (SES) and language ability across domains in early childhood. We then identify three potential pathways by which SES might influence language development—child characteristics, parent–child interaction, and availability of learning resources—recognizing the complicated interaction between the child's own language learning skill and his/her environmental support. Finally, we review interventions that target these three pathways with an eye toward best practice. Future research should focus on the diversity of contexts in which children acquire language and adopt methods of language measurement that are sensitive to cultural variation.

This review addresses several situations of language learning to make concrete the issue of fairness—and justice—that arises in designing assessments. First, I discuss the implications of dialect variation in American English, asking how assessment has taken dialect into consideration. Second, I address the question of how to assess the distributed knowledge of bilingual or dual-language learners. The evaluation of the language skills of children growing up in poverty asks whether the current focus on the quantity of caregiver input is misplaced. Third, I address a special case in which the young speakers of a minority language, Romani, are judged to be unfit for schooling because they fail tests in the state language. Finally, I examine the difficult issue of language assessments in countries with multiple official languages and few resources. In each of these areas, the involvement and expertise of linguists are essential for knowing how the grammar works and what might be important to assess.

A great deal of research during the past 40 years and earlier investigates the outcomes of large-scale contact between mutually intelligible dialects. In the context of well-established types of dialect convergence, this article summarizes research from the past 10 years on migration-induced dialect contact. I devote particular attention to ongoing dialect contact in the urban settings of London, England; São Paulo, Brazil; Xining, China; Amman, Jordan; and New York City, United States. In all cases, the data show evidence of the expected process of dialect leveling, but linguistic innovations emerging from dialect contact are also prominent. Social network and identity factors are predicted to define much future dialect contact research.

This article explores major processes that operate in new dialect formation, with a focus on the effects of various kinds of isolation of communities and their speakers. A sociolinguistic approach concentrates on competition and selection between transplanted dialect features and mechanisms such as mixing, leveling, simplification, and reallocation. With reference to the established (yet not uncontroversial) concept of the feature pool, the question is how contact between distinct systems gives rise to localized forms. From a more socially oriented perspective, the question is how the integration and segregation of groups contributes to or shapes the emergence and disappearance of dialects. Isolation as a sociolinguistic concept is not thoroughly defined, particularly at the intergroup, societal, or even national level; therefore, geographic, social, and psychological factors need to be discussed and assessed. The processes investigated here are koinéization and dialect/language shift, illustrated with examples from all over the world.

Language emergence describes moments in historical time when nonlinguistic systems become linguistic. Because language can be invented de novo in the manual modality, this offers insight into the emergence of language in ways that the oral modality cannot. Here we focus on homesign, gestures developed by deaf individuals who cannot acquire spoken language and have not been exposed to sign language. We contrast homesign with (a) gestures that hearing individuals produce when they speak, as these cospeech gestures are a potential source of input to homesigners, and (b) established sign languages, as these codified systems display the linguistic structure that homesign has the potential to assume. We find that the manual modality takes on linguistic properties, even in the hands of a child not exposed to a language model. But it grows into full-blown language only with the support of a community that transmits the system to the next generation.*

Why are languages the way they are? The biases and constraints that explain why languages display the traits they do—instead of other possible ones—include human cognition, social dynamics, communicative function, the structure of meanings, and the interactions between these. Cultural language evolution is concerned with explaining the causal pathways that link these biases and constraints with fundamental linguistic traits, such as combinatoriality, compositionality, conventionality, and arbitrariness. The cultural evolution of language, or the emergence of a language from no language, begins when motivated signals are used in context by individuals who do not share a means of communication. Repeated interaction between interlocutors makes individual signals compressible through reduction or simplification and can entrench idiosyncratic patterns. Transmission to new learners results in the increasing compressibility at the level of the system through the introduction of categories and regular rules, and in the spread of conventional, arbitrary signals.