Human experience of color results from a complex interplay of perceptual and linguistic systems. At the lowest level of perception, the human visual system transforms the visible light portion of the electromagnetic spectrum into a rich, continuous three-dimensional experience of color. Despite our ability to perceptually discriminate millions of different color shades, most languages categorize color into a number of discrete color categories. While the meanings of color words are constrained by perception, perception does not fully define them. Once color words are acquired, they may in turn influence our memory and processing speed for color, although it is unlikely that language influences the lowest levels of color perception.
One approach to examining the relationship between perception and language in forming our experience of color is to study children as they acquire color language. Children produce color words in speech for many months before acquiring adult meanings for color words. Research in this area has focused on whether children’s difficulties stem from (a) an inability to identify color properties as a likely candidate for word meanings, or alternatively (b) inductive learning of language-specific color word boundaries. Lending plausibility to the first account, there is evidence that children more readily attend to object traits like shape, rather than color, as likely candidates for word meanings. However, recent evidence has found that children have meanings for some color words before they begin to produce them in speech, indicating that in fact, they may be able to successfully identify color as a candidate for word meaning early in the color word learning process. There is also evidence that prelinguistic infants, like adults, perceive color categorically. While these perceptual categories likely constrain the meanings that children consider, they cannot fully define color word meanings because languages vary in both the number and location of color word boundaries. Recent evidence suggests that the delay in color word acquisition primarily stems from an inductive process of refining these boundaries.
Article
The Acquisition of Color Words
Katie Wagner and David Barner
Article
The Acquisition of Word-Formation in the Romance Languages
Eve V. Clark
Several factors influence children’s initial choices of word-formation options––simplicity of form, transparency of meaning, and productivity in current adult speech. The coining of new words is also constrained by general pragmatic considerations for usage: Reliance on conventionality, contrast, and cooperation between speaker and addressee. For children acquiring French, Italian, Portuguese, and Spanish, the data on what they know about word-formation for the coining of new words consist primarily of diary observations; in some cases, these are supplemented with experimental elicitation studies of the comprehension and production of new word-forms. The general patterns in Romance acquisition of word-formation favor derivation over compounding. Children produce some spontaneous coinages with zero derivation (verbs converted to nouns in French, for example) from as young as 2 years, 6 months (2;6). The earliest suffixes children put to use in these languages tend to be agentive (from 2;6 to 3 years onward), followed by instrumental, objective, locative, and, slightly later, diminutive. The only prefixes that emerge early in child innovations are negative ones used to express reversals of actions. Overall, the general patterns of acquisition for word-formation in Romance are similar to those in Semitic, where derivation is also more productive than compounding, rather than to those in Germanic, where compounding is highly productive, and emerges very early, before any derivational forms.
Article
Affixation in Morphology
Kristel Van Goethem
Affixation is the morphological process that consists of adding an affix (i.e., a bound morpheme) to a morphological base. It is cross-linguistically the most common process that human languages use to derive new lexemes (derivational affixation) or to adapt a word’s form to its morphosyntactic context (inflectional affixation). Suffixes (i.e., bound morphemes following the base) and prefixes (i.e., bound morphemes preceding the base) are the most common affixes, with suffixation being more frequently recorded in the world’s languages than prefixation. Minor types of affixation include circumfixation and infixation. Conversion and back-formation are related derivational processes that do not make use of affixation.
Many studies have concentrated on the need to differentiate derivation from inflection, but these morphological processes are probably best described as two end points of a cline. Prototypically, derivation is used to change a word’s category (part of speech) and involves a semantic change. A word’s inflectional distinctions make up its paradigm, which amounts to the different morphological forms that correlate with different morphosyntactic functions. Form-function mapping in (derivational and inflectional) affixation is a key issue in current research on affixation. Many deviations from the canonical One Form-One Meaning principle can be observed in the field of affixation.
From a diachronic point of view, it has been demonstrated that affixes often derive from free lexemes by grammaticalization, with affixoids being recognized as an intermediate step on this cline. More controversial, but still attested, is the opposite change whereby affixes and affixoids develop into free morphemes through a process of degrammaticalization.
Article
Alignment and Word Order in the Romance Languages
Francesco Rovai
The term “alignment” refers to the formal realization of the argument structure of the clause, that is, the ways in which the core arguments of the predicate are encoded by means of three main morphosyntactic devices: nominal case marking (morphological case, adpositions), verb marking systems (verbal agreement, pronominal affixes, auxiliaries, voice distinctions, etc.), and word order. The relative importance of these mechanisms of argument coding may considerably vary from language to language. In the Romance family, a major role is played by finite verb agreement and, to a lesser extent, auxiliary selection, participial agreement, voice distinctions, and word order, depending on the language/variety. Most typically, both transitive and intransitive subjects share the same formal coding (they control finite verb agreement and precede the verb in the basic word order) and are distinguished from direct objects (which do not control finite verb agreement and follow the verb in the basic word order). This arrangement of the argument structure is traditionally known as “nominative/accusative” alignment and can be easily identified as the main alignment of the Romance languages. Note that, with very few exceptions, nominal case marking is instead “neutral,” since no overt morphological distinction is made between subject and object arguments after the loss of the Latin case system.
However, although the Romance languages can legitimately be associated with an accusative alignment, it must be borne in mind that, whatever the property selected, natural languages speak against an all-encompassing, holistic typology. A language “belongs” to an alignment type only insofar as it displays a significantly above-average frequency of clause structures with that kind of argument coding, but this does not exclude the existence of several grammatical domains that partake of different alignments. In the Romance family, minor patterns are attested that are not consistent with an accusative alignment. In part, they depend on robust crosslinguistic tendencies in the distribution of the different alignment types when they coexist in the same language. In part, they reflect phenomena of morphosyntactic realignment that can be traced back to the transition from Latin to Romance, when, alongside the dominant accusative alignment of the classical language, Late Latin developed an active alignment in some domains of the grammar—a development that has its roots in Classical and Early Latin. Today, the Romance languages preserve traces of this intermediate stage, but in large part, the signs of it have been replaced with novel accusative structures. In particular, at the level of the sentence, there emerges an accusative-aligned word order, with the preverbal position realizing the default “subject” position and the postverbal position instantiating the default “object” position.
Article
Argument Realization and Case in Japanese
Hideki Kishimoto
Japanese is a language where the grammatical status of arguments and adjuncts is marked exclusively by postnominal case markers, and various argument realization patterns can be assessed by their case marking. Since Japanese is categorized as a language of the nominative-accusative type typologically, the unmarked case-marking frame obtained for transitive predicates of the non-stative (or eventive) type is ‘nominative-accusative’. Nevertheless, transitive predicates falling into the stative class often have other case-marking alignments, such as ‘nominative-nominative’ and ‘dative-nominative’. Consequently, Japanese provides much more varying argument realization patterns than those expected from its typological character as a nominative-accusative language.
In point of fact, argument marking can actually be much more elastic and variable, the variations being motivated by several linguistic factors. Arguments often have the option of receiving either syntactic or semantic case, with no difference in the logical or cognitive meaning (as in plural agent and source agent alternations) or depending on the meanings their predicate carry (as in locative alternation). The type of case marking that is not normally available in main clauses can sometimes be obtained in embedded contexts (i.e., in exceptional case marking and small-clause constructions). In complex predicates, including causative and indirect passive predicates, arguments are case-marked differently from their base clauses by virtue of suffixation, and their case patterns follow the mono-clausal case array, despite the fact that they have multi-clausal structures.
Various case marking options are also made available for arguments by grammatical operations. Some processes instantiate a change on the grammatical relations and case marking of arguments with no affixation or embedding. Japanese has the grammatical process of subjectivization, creating extra (non-thematic) major subjects, many of which are identified as instances of ‘possessor raising’ (or argument ascension). There is another type of grammatical process, which reduces the number of arguments by virtue of incorporating a noun into the predicate, as found in the light verb constructions with suru ‘do’ and the complex adjective constructions formed on the negative adjective nai ‘non-existent.’
Article
Argument Realization in Syntax
Malka Rappaport Hovav
Words are sensitive to syntactic context. Argument realization is the study of the relation between argument-taking words, the syntactic contexts they appear in and the interpretive properties that constrain the relation between them.
Article
Attributive Compounds
Anton Karl Ingason and Einar Freyr Sigurðsson
Attributive compounds are words that include two parts, a head and a non-head, both of which include lexical roots, and in which the non-head is interpreted as a modifier of the head. The nature of this modification is sometimes described in terms of a covert relationship R. The nature of R has been the subject of much discussion in the literature, including proposals that a finite and limited number of interpretive options are available for R, as well as analyses in which the interpretation of R is unrestricted and varies with context. The modification relationship between the parts of an attributive compound also contrasts with the interpretation of compounds in other ways because some non-heads in compounds saturate argument positions of the head, others are semantically conjoined with them, and some restrict their domain of interpretation.
Article
Blending in Morphology
Natalia Beliaeva
Blending is a type of word formation in which two or more words are merged into one so that the blended constituents are either clipped, or partially overlap. An example of a typical blend is brunch, in which the beginning of the word breakfast is joined with the ending of the word lunch. In many cases such as motel (motor + hotel) or blizzaster (blizzard + disaster) the constituents of a blend overlap at segments that are phonologically or graphically identical. In some blends, both constituents retain their form as a result of overlap, for example, stoption (stop + option). These examples illustrate only a handful of the variety of forms blends may take; more exotic examples include formations like Thankshallowistmas (Thanksgiving + Halloween + Christmas). The visual and audial amalgamation in blends is reflected on the semantic level. It is common to form blends meaning a combination or a product of two objects or phenomena, such as an animal breed (e.g., zorse, a breed of zebra and horse), an interlanguage variety (e.g., franglais, which is a French blend of français and anglais meaning a mixture of French and English languages), or other type of mix (e.g., a shress is a type of clothes having features of both a shirt and a dress).
Blending as a word formation process can be regarded as a subtype of compounding because, like compounds, blends are formed of two (or sometimes more) content words and semantically either are hyponyms of one of their constituents, or exhibit some kind of paradigmatic relationships between the constituents. In contrast to compounds, however, the formation of blends is restricted by a number of phonological constraints given that the resulting formation is a single word. In particular, blends tend to be of the same length as the longest of their constituent words, and to preserve the main stress of one of their constituents. Certain regularities are also observed in terms of ordering of the words in a blend (e.g., shorter first, more frequent first), and in the position of the switch point, that is, where one blended word is cut off and switched to another (typically at the syllable boundary or at the onset/rime boundary). The regularities of blend formation can be related to the recognizability of the blended words.
Article
Bracketing Paradoxes in Morphology
Heather Newell
Bracketing paradoxes—constructions whose morphosyntactic and morpho-phonological structures appear to be irreconcilably at odds (e.g., unhappier)—are unanimously taken to point to truths about the derivational system that we have not yet grasped. Consider that the prefix un- must be structurally separate in some way from happier both for its own reasons (its [n] surprisingly does not assimilate in Place to a following consonant (e.g., u[n]popular)), and for reasons external to the prefix (the suffix -er must be insensitive to the presence of un-, as the comparative cannot attach to bases of three syllables or longer (e.g., *intelligenter)). But, un- must simultaneously be present in the derivation before -er is merged, so that unhappier can have the proper semantic reading (‘more unhappy’, and not ‘not happier’). Bracketing paradoxes emerged as a problem for generative accounts of both morphosyntax and morphophonology only in the 1970s. With the rise of restrictions on and technology used to describe and represent the behavior of affixes (e.g., the Affix-Ordering Generalization, Lexical Phonology and Morphology, the Prosodic Hierarchy), morphosyntacticians and phonologists were confronted with this type of inconsistent derivation in many unrelated languages.
Article
Chinese Dou Quantification
Yuli Feng and Haihua Pan
Dou has been seen as a typical example of universal quantification and the point of departure in the formal study of quantification in Chinese. The constraints on dou’s quantificational structure, dou’s diverse uses, and the compatibility between dou and other quantificational expressions have further promoted the refinement of the theory of quantification and sparked debate over the semantic nature of dou. The universal quantificational approach holds that dou is a universal quantifier and explains its diverse uses as the effects produced by quantification on different sorts of entities and different ways of quantificational mapping. However, non-quantificational approaches, integrating the insights of degree semantics and focus semantics, take the scalar use as dou’s core semantics. The quantificational approach to dou can account for its meaning of exclusiveness and the interpretational differences engendered by dou when it associates with a wh-indeterminate to its left or to its right, whereas non-quantificational approaches cannot determine the interpretational differences caused by rightward and leftward association and cannot explain the exclusive use of dou. Despite the differences, the various approaches to dou, quantificational or non-quantificational, have far-reaching theoretical significance for understanding the mechanism of quantification in natural language.
Article
Liheci ‘Separable Words’ in Mandarin Chinese
Kuang Ye and Haihua Pan
Liheci ‘separable words’ is a special phenomenon in Mandarin Chinese, and it refers to an intransitive verb with two or more syllables that allows the insertion of syntactic modifiers or an argument in between the first syllable and the second or the rest of syllables with the help of the nominal modifier marker de. There are two major groups of Liheci: those stored in the lexicon, such as bangmang ‘help’, lifa ‘haircut’, and shenqi ‘anger’, and those derived in syntax through noun-to-verb incorporation, such as chifan ‘eat meal’, leiqiang ‘build wall’, in which fan ‘meal’ and qiang ‘wall’ are incorporated into chi ‘eat’ and lei ‘build’, respectively, to function as temporary verbal compounds. The well-known behavior of Liheci is that it can be separated by nominal modifiers or a syntactic argument. For example, bangmang ‘help’ can be used to form a verb phrase bang Lisi-de mang ‘give Lisi a help’ by inserting Lisi and a nominal modifier marker, de, between bang and mang, with bang being understood as the predicate and Lisi-de mang as the object. Although Lisi appears as a possessor marked by de, it should be understood as the theme object of the compound verb. In similar ways, the syntactic–semantic elements such as agent, theme, adjectives, measure phrases, relative clauses, and the like can all be inserted between the two components of bangmang, deriving verb phrases like (Zhangsan) bang Zhangsan-de mang ‘(Zhangsan) do Zhangsan’s help’, where Zhangsan is the agent; bang-le yi-ci mang ‘help once’, where yi-ci is a measure phrase; and bang bieren bu xiang bang de mang ‘give a help that others don’t want to give’, where bieren bu xiang bang is a relative clause. The same insertions can be found in Liheci formed in syntax. For example, chi liang-ci fan ‘eat two time’s meal’ (eat meals twice), lei san-tian qiang ‘build three day’s wall’ (build walls for three days).
There are three syntactic-semantic properties exhibited in verb phrases formed with Liheci: first, possessors being understood as Liheci’s logical argument; second, interdependent relation between the predicate and the complement; and, third, obligatory use of verbal classifiers instead of nominal classifiers.
In this article, first, five influential analyses in the literature are reviewed, pointing out their strengths and weaknesses. Then, the cognate object approach is discussed. Under this approach, Lihecis are found to be intransitive verbs that are capable of taking nominalized reduplicates of themselves as their cognate objects. After a complementary deletion on the verb and its reduplicate object in the Phonetic Form (PF), all the relevant verb phrases can be well derived, with no true separation involved in the derivation, as all the copies of Liheci in question remain intact all along. After a discussion of the relevant syntactic structures, it is shown that with this syntactic capacity, all participants involved in the events can be successfully accommodated and correctly interpreted. The advantage can be manifested in six aspects, demonstrating that this proposal fares much better than other approaches.
Article
Chinese Semantics
Haihua Pan and Yuli Feng
Cross-linguistic data can add new insights to the development of semantic theories or even induce the shift of the research paradigm. The major topics in semantic studies such as bare noun denotation, quantification, degree semantics, polarity items, donkey anaphora and binding principles, long-distance reflexives, negation, tense and aspects, eventuality are all discussed by semanticists working on the Chinese language. The issues which are of particular interest include and are not limited to: (i) the denotation of Chinese bare nouns; (ii) categorization and quantificational mapping strategies of Chinese quantifier expressions (i.e., whether the behaviors of Chinese quantifier expressions fit into the dichotomy of A-Quantification and D-quantification); (iii) multiple uses of quantifier expressions (e.g., dou) and their implication on the inter-relation of semantic concepts like distributivity, scalarity, exclusiveness, exhaustivity, maximality, etc.; (iv) the interaction among universal adverbials and that between universal adverbials and various types of noun phrases, which may pose a challenge to the Principle of Compositionality; (v) the semantics of degree expressions in Chinese; (vi) the non-interrogative uses of wh-phrases in Chinese and their influence on the theories of polarity items, free choice items, and epistemic indefinites; (vii) how the concepts of E-type pronouns and D-type pronouns are manifested in the Chinese language and whether such pronoun interpretations correspond to specific sentence types; (viii) what devices Chinese adopts to locate time (i.e., does tense interpretation correspond to certain syntactic projections or it is solely determined by semantic information and pragmatic reasoning); (ix) how the interpretation of Chinese aspect markers can be captured by event structures, possible world semantics, and quantification; (x) how the long-distance binding of Chinese ziji ‘self’ and the blocking effect by first and second person pronouns can be accounted for by the existing theories of beliefs, attitude reports, and logophoricity; (xi) the distribution of various negation markers and their correspondence to the semantic properties of predicates with which they are combined; and (xii) whether Chinese topic-comment structures are constrained by both semantic and pragmatic factors or syntactic factors only.
Article
Chinese Verbs and Lexical Distinction
Meichun Liu
Chinese verbs behave very differently from their counterparts in Indo-European languages and pose interesting challenges to the study of syntax-semantic interface for theoretical and applicational linguistics. The lexical semantic distinctions encoded in the Chinese verbal lexicon are introduced with a thorough review of previous works from different approaches with different concerns and answers. The recent development in constructing a digital database of verbal information in Mandarin Chinese, the Mandarin VerbNet, is also introduced, which offers frame-based constructional analyses of the Chinese verbs and verb classes. Finally, a case study on Chinese emotion verbs is presented to illustrate the unique properties of lexicalization patterns in Chinese verbs. In general, due to its typological characteristics in coding a Topic, rather than a Subject, as a prominent element in the sentence, Chinese shows a more flexible range of form-meaning mapping relations in lexical distinctions.
Article
Clinical Linguistics
Louise Cummings
Clinical linguistics is the branch of linguistics that applies linguistic concepts and theories to the study of language disorders. As the name suggests, clinical linguistics is a dual-facing discipline. Although the conceptual roots of this field are in linguistics, its domain of application is the vast array of clinical disorders that may compromise the use and understanding of language. Both dimensions of clinical linguistics can be addressed through an examination of specific linguistic deficits in individuals with neurodevelopmental disorders, craniofacial anomalies, adult-onset neurological impairments, psychiatric disorders, and neurodegenerative disorders. Clinical linguists are interested in the full range of linguistic deficits in these conditions, including phonetic deficits of children with cleft lip and palate, morphosyntactic errors in children with specific language impairment, and pragmatic language impairments in adults with schizophrenia.
Like many applied disciplines in linguistics, clinical linguistics sits at the intersection of a number of areas. The relationship of clinical linguistics to the study of communication disorders and to speech-language pathology (speech and language therapy in the United Kingdom) are two particularly important points of intersection. Speech-language pathology is the area of clinical practice that assesses and treats children and adults with communication disorders. All language disorders restrict an individual’s ability to communicate freely with others in a range of contexts and settings. So language disorders are first and foremost communication disorders. To understand language disorders, it is useful to think of them in terms of points of breakdown on a communication cycle that tracks the progress of a linguistic utterance from its conception in the mind of a speaker to its comprehension by a hearer. This cycle permits the introduction of a number of important distinctions in language pathology, such as the distinction between a receptive and an expressive language disorder, and between a developmental and an acquired language disorder. The cycle is also a useful model with which to conceptualize a range of communication disorders other than language disorders. These other disorders, which include hearing, voice, and fluency disorders, are also relevant to clinical linguistics.
Clinical linguistics draws on the conceptual resources of the full range of linguistic disciplines to describe and explain language disorders. These disciplines include phonetics, phonology, morphology, syntax, semantics, pragmatics, and discourse. Each of these linguistic disciplines contributes concepts and theories that can shed light on the nature of language disorder. A wide range of tools and approaches are used by clinical linguists and speech-language pathologists to assess, diagnose, and treat language disorders. They include the use of standardized and norm-referenced tests, communication checklists and profiles (some administered by clinicians, others by parents, teachers, and caregivers), and qualitative methods such as conversation analysis and discourse analysis. Finally, clinical linguists can contribute to debates about the nosology of language disorders. In order to do so, however, they must have an understanding of the place of language disorders in internationally recognized classification systems such as the 2013 Diagnostic and Statistical Manual of Mental Disorders (DSM-5) of the American Psychiatric Association.
Article
Cognitive Semantics in the Romance Languages
Ulrich Detges
Cognitive semantics (CS) is an approach to the study of linguistic meaning. It is based on the assumption that the human linguistic capacity is part of our cognitive abilities, and that language in general and meaning in particular can therefore be better understood by taking into account the cognitive mechanisms that control the conceptual and perceptual processing of extra-linguistic reality. Issues central to CS are (a) the notion of prototype and its role in the description of language, (b) the nature of linguistic meaning, and (c) the functioning of different types of semantic relations. The question concerning the nature of meaning is an issue that is particularly controversial between CS on the one hand and structuralist and generative approaches on the other hand: is linguistic meaning conceptual, that is, part of our encyclopedic knowledge (as is claimed by CS), or is it autonomous, that is, based on abstract and language-specific features? According to CS, the most important types of semantic relations are metaphor, metonymy, and different kinds of taxonomic relations, which, in turn, can be further broken down into more basic associative relations such as similarity, contiguity, and contrast. These play a central role not only in polysemy and word formation, that is, in the lexicon, but also in the grammar.
Article
Collectives in the Romance Languages
Wiltrud Mihatsch
Just like other semantic subtypes of nouns such as event nouns or agent nouns, collectives may be morphologically opaque lexemes, but they are also regularly derived in many languages. Perhaps not a word-formation category as productive as event nouns or agent nouns, collective nouns still represent a category associated with particular means of word formation, in the case of the Romance languages by means of derivational suffixes. The Romance languages all have suffixes for deriving collectives, but only very few go directly back to Latin. In most cases, they evolve from other derivational suffixes via metonymic changes of individual derived nouns, notably event nouns and quality nouns. Due to the ubiquity of these changes, series of semantically and morphologically equivalent collectives trigger functional changes of the suffixes themselves, which may then acquire collective meaning. Most of these suffixes are pan-Romance, in many cases going back to very early changes, or to inter-Romance loans. The different Romance languages have overlapping inventories of suffixes, with different degrees of productivity and different semantic niches. The ease of transition from event or quality noun to collective also explains why only few suffixes are exclusively used for the derivation of collective nouns.
Article
The Compositional Semantics of Modification
Sebastian Bücking
Modification is a combinatorial semantic operation between a modifier and a modifiee. Take, for example, vegetarian soup: the attributive adjective vegetarian modifies the nominal modifiee soup and thus constrains the range of potential referents of the complex expression to soups that are vegetarian. Similarly, in Ben is preparing a soup in the camper, the adverbial in the camper modifies the preparation by locating it. Notably, modifiers can have fairly drastic effects; in fake stove, the attribute fake induces that the complex expression singles out objects that seem to be stoves, but are not. Intuitively, modifiers contribute additional information that is not explicitly called for by the target the modifier relates to. Speaking in terms of logic, this roughly says that modification is an endotypical operation; that is, it does not change the arity, or logical type, of the modified target constituent. Speaking in terms of syntax, this predicts that modifiers are typically adjuncts and thus do not change the syntactic distribution of their respective target; therefore, modifiers can be easily iterated (see, for instance, spicy vegetarian soup or Ben prepared a soup in the camper yesterday). This initial characterization sets modification apart from other combinatorial operations such as argument satisfaction and quantification: combining a soup with prepare satisfies an argument slot of the verbal head and thus reduces its arity (see, for instance, *prepare a soup a quiche). Quantification as, for example, in the combination of the quantifier every with the noun soup, maps a nominal property onto a quantifying expression with a different distribution (see, for instance, *a every soup). Their comparatively loose connection to their hosts renders modifiers a flexible, though certainly not random, means within combinatorial meaning constitution. The foundational question is how to work their being endotypical into a full-fledged compositional analysis. On the one hand, modifiers can be considered endotypical functors by virtue of their lexical endowment; for instance, vegetarian would be born a higher-ordered function from predicates to predicates. On the other hand, modification can be considered a rule-based operation; for instance, vegetarian would denote a simple predicate from entities to truth-values that receives its modifying endotypical function only by virtue of a separate modification rule. In order to assess this and related controversies empirically, research on modification pays particular attention to interface questions such as the following: how do structural conditions and the modifying function conspire in establishing complex interpretations? What roles do ontological information and fine-grained conceptual knowledge play in the course of concept combination?
Article
Compound and Complex Predicates in Japanese
Taro Kageyama
Compound and complex predicates—predicates that consist of two or more lexical items and function as the predicate of a single sentence—present an important class of linguistic objects that pertain to an enormously wide range of issues in the interactions of morphology, phonology, syntax, and semantics. Japanese makes extensive use of compounding to expand a single verb into a complex one. These compounding processes range over multiple modules of the grammatical system, thus straddling the borders between morphology, syntax, phonology, and semantics. In terms of degree of phonological integration, two types of compound predicates can be distinguished. In the first type, called tight compound predicates, two elements from the native lexical stratum are tightly fused and inflect as a whole for tense. In this group, Verb-Verb compound verbs such as arai-nagasu [wash-let.flow] ‘to wash away’ and hare-agaru [sky.be.clear-go.up] ‘for the sky to clear up entirely’ are preponderant in numbers and productivity over Noun-Verb compound verbs such as tema-doru [time-take] ‘to take a lot of time (to finish).’
The second type, called loose compound predicates, takes the form of “Noun + Predicate (Verbal Noun [VN] or Adjectival Noun [AN]),” as in post-syntactic compounds like [sinsya : koonyuu] no okyakusama ([new.car : purchase] GEN customers) ‘customer(s) who purchase(d) a new car,’ where the symbol “:” stands for a short phonological break. Remarkably, loose compounding allows combinations of a transitive VN with its agent subject (external argument), as in [Supirubaagu : seisaku] no eiga ([Spielberg : produce] GEN film) ‘a film/films that Spielberg produces/produced’—a pattern that is illegitimate in tight compounds and has in fact been considered universally impossible in the world’s languages in verbal compounding and noun incorporation.
In addition to a huge variety of tight and loose compound predicates, Japanese has an additional class of syntactic constructions that as a whole function as complex predicates. Typical examples are the light verb construction, where a clause headed by a VN is followed by the light verb suru ‘do,’ as in Tomodati wa sinsya o koonyuu (sae) sita [friend TOP new.car ACC purchase (even) did] ‘My friend (even) bought a new car’ and the human physical attribute construction, as in Sensei wa aoi me o site-iru [teacher TOP blue eye ACC do-ing] ‘My teacher has blue eyes.’ In these constructions, the nominal phrases immediately preceding the verb suru are semantically characterized as indefinite and non-referential and reject syntactic operations such as movement and deletion. The semantic indefiniteness and syntactic immobility of the NPs involved are also observed with a construction composed of a human subject and the verb aru ‘be,’ as Gakkai ni wa oozei no sankasya ga atta ‘There was a large number of participants at the conference.’ The constellation of such “word-like” properties shared by these compound and complex predicates poses challenging problems for current theories of morphology-syntax-semantics interactions with regard to such topics as lexical integrity, morphological compounding, syntactic incorporation, semantic incorporation, pseudo-incorporation, and indefinite/non-referential NPs.
Article
Computational Semantics
Katrin Erk
Computational semantics performs automatic meaning analysis of natural language. Research in computational semantics designs meaning representations and develops mechanisms for automatically assigning those representations and reasoning over them. Computational semantics is not a single monolithic task but consists of many subtasks, including word sense disambiguation, multi-word expression analysis, semantic role labeling, the construction of sentence semantic structure, coreference resolution, and the automatic induction of semantic information from data.
The development of manually constructed resources has been vastly important in driving the field forward. Examples include WordNet, PropBank, FrameNet, VerbNet, and TimeBank. These resources specify the linguistic structures to be targeted in automatic analysis, and they provide high-quality human-generated data that can be used to train machine learning systems. Supervised machine learning based on manually constructed resources is a widely used technique.
A second core strand has been the induction of lexical knowledge from text data. For example, words can be represented through the contexts in which they appear (called distributional vectors or embeddings), such that semantically similar words have similar representations. Or semantic relations between words can be inferred from patterns of words that link them. Wide-coverage semantic analysis always needs more data, both lexical knowledge and world knowledge, and automatic induction at least alleviates the problem.
Compositionality is a third core theme: the systematic construction of structural meaning representations of larger expressions from the meaning representations of their parts. The representations typically use logics of varying expressivity, which makes them well suited to performing automatic inferences with theorem provers.
Manual specification and automatic acquisition of knowledge are closely intertwined. Manually created resources are automatically extended or merged. The automatic induction of semantic information is guided and constrained by manually specified information, which is much more reliable. And for restricted domains, the construction of logical representations is learned from data.
It is at the intersection of manual specification and machine learning that some of the current larger questions of computational semantics are located. For instance, should we build general-purpose semantic representations, or is lexical knowledge simply too domain-specific, and would we be better off learning task-specific representations every time? When performing inference, is it more beneficial to have the solid ground of a human-generated ontology, or is it better to reason directly with text snippets for more fine-grained and gradual inference? Do we obtain a better and deeper semantic analysis as we use better and deeper manually specified linguistic knowledge, or is the future in powerful learning paradigms that learn to carry out an entire task from natural language input and output alone, without pre-specified linguistic knowledge?
Article
Connectionism in Linguistic Theory
Xiaowei Zhao
Connectionism is an important theoretical framework for the study of human cognition and behavior. Also known as Parallel Distributed Processing (PDP) or Artificial Neural Networks (ANN), connectionism advocates that learning, representation, and processing of information in mind are parallel, distributed, and interactive in nature. It argues for the emergence of human cognition as the outcome of large networks of interactive processing units operating simultaneously. Inspired by findings from neural science and artificial intelligence, connectionism is a powerful computational tool, and it has had profound impact on many areas of research, including linguistics. Since the beginning of connectionism, many connectionist models have been developed to account for a wide range of important linguistic phenomena observed in monolingual research, such as speech perception, speech production, semantic representation, and early lexical development in children. Recently, the application of connectionism to bilingual research has also gathered momentum. Connectionist models are often precise in the specification of modeling parameters and flexible in the manipulation of relevant variables in the model to address relevant theoretical questions, therefore they can provide significant advantages in testing mechanisms underlying language processes.