the relative importance attributed to form and structure vs. function and context- dependence as criterial properties of language (Section 2.1.2); the related focus on competence vs. performance (Section 2.1.3); the relative continuity vs. discontinuity in the course of acquisition (2.1.4); different views of the relation between language and cognition during development, varying in terms of whether and how language may have a structuring role on thought (Section 2.1.5). Although I highlight these issues by sharply contrasting views for ease of presentation, it should be clear that at least some of these controversies might be best viewed as reflecting dimensions or continua, along which theories can be placed as a function of their primary focus.
2.1.1 Nativistic and constructivistic approaches
The first debate, and perhaps the most fundamental one, concerns the relative role of innate knowledge vs. gradual learning mechanisms in our account
Theoretical issues
of language development. I contrast here ‘nativistic’ vs. ‘constructivistic’ views, focusing on three points around which these views differ. First, these different positions have focused on different aspects of the timing of acquisition, choosing to highlight either precocious knowledge or gradual and/or late developments. Second, and most importantly, these different viewpoints make different assumptions about children’s biological endowment at birth, postulating different degrees of initial knowledge to account for ontogenesis. Finally, the answers proposed to this second question have implications for the ways in which models go about accounting for observed developmental changes, postulating very different mechanisms for ontogenetic development.
2.1.1.1 Timing
With respect to timing in acquisition, some approaches assume that language is discovered instantaneously or extremely early by the child, whereas others assume that it is acquired during a long and gradual process, resulting in rather late developments. Innateness has been typically invoked in the face of the striking speed and systematicity with which all children seem to acquire their native language. Regardless of the wide diversity of environmental factors surrounding them, and despite an imperfect and fragmentary input, children seem quickly to dis- play knowledge of the rules of grammar, allowing them to produce and comprehend an apparently infinite number of well-formed sentences. This phenomenon has led some to postulate that language acquisition is determined by biological factors spe- cific to our species and to predict strikingly early developments. Researchers in this tradition have indeed pointed out surprisingly precocious language abilities from the initial phase characterised by the emergence of productive speech (at around eighteen months) to a later phase (typically said to occur before five years), where children have been assumed to master language, particularly its grammatical sys- tem. Some recent research (e.g. Hirsh-Pasek and Golinkoff 1996) has also aimed at showing that yet more precocious recognition or comprehension processes occur before the initial phase of emergence.
In contrast, gradual learning processes have been typically invoked in the face of two types of phenomena. First, some striking regularities can be observed during the process of language acquisition, resulting in recurrent developmental sequences.
As shown below, such regularities have been attributed to various factors, either endogenous cognitive factors or exogenous social factors related to interpersonal interaction. Second, some differences can also be observed in these sequences across languages, which have been attributed to the impact of language-specific or culture- specific factors. Indeed, some researchers within the constructivistic approach have argued that many developments occur after five years (until at least twelve years),
Some issues in theories of language acquisition including the mastery of syntactic, semantic, and pragmatic aspects of language necessary for the organisation of discourse units that go beyond the sentence. These late developments have been typically attributed to external ‘performance factors’
within the first approach, such as factors linked to contextual constraints or to the limits of the child’s developing cognitive system, which are claimed to be central to the child’s linguistic competence within the second approach. I return to this point below.
2.1.1.2 Biological endowment in ontogenesis
A fundamental point of contrast among theories of language acquisition concerns the type of biological endowment they attribute to human infants at birth.
Although early versions of this question asked whether some endowment should be acknowledged at all, this controversy has evolved, revolving now around the nature and specificity of this endowment. My purpose here is merely to point out some aspects of this complex question, which is central to several types of accounts concerning phylogenetic and ontogenetic development. On a phylogenetic scale, various debates concern the types of capacities that might be either entirely specific to our species or only found in much less developed forms in other species of the animal world. Among the capacities that have been claimed to be human-specific are abstract conceptual organisation and logico-mathematical reasoning, some forms of symbolic behaviour, the capacity for self-reflexive behaviour and more generally for higher forms of consciousness, complex interpersonal communication and social organisation, and special types of perceptual organisation (see Langer 2001 for some counter-arguments).
One aspect of this issue is the initial equipment that has been attributed to children at birth. Two extreme positions can be contrasted for the sake of presentation:nativi- sticandbehaviouristviews. At one extreme, nativistic models postulate that infants are equipped at birth with a brain that is entirely pre-programmed for the structure of human language. At the other extreme, behaviourist models postulate that the infant is at first atabula rasathat is not endowed with any pre-programmed capacities. Each of these theoretical poles comprises weak and strong versions, depending on the extent to which they allow for qualifications to their main assumptions. In addition, other types of positions have also been proposed between these two poles, varying with respect to other dimensions. Among them, at least two major orientations can be distinguished:cognitivistandinteractionistmodels. The former type of model postulates a general cognitive capacity allowing the infant to construct a gradually more complex representation of the world, while the latter type equips it with an initial capacity for social interaction, allowing it to participate in gradually more complex forms of communication. In both cases, language plays an important role
Theoretical issues
in providing a powerful symbolic system for the child’s epistemological system or for its interpersonal interactions, both of which might be specific to the human species. However, as will be shown, theories vary here a great deal with respect to how special or crucial they consider the role of language to be and/or in whether or not they view particular aspects of children’s cognitive or communicative capacities as being pre-programmed and species-specific.
The issue of innateness has come back to the forefront of various cognitive sciences, given a great number of accumulated results concerning the capacities of young infants in various domains, which have been uncovered since the middle of the twentieth century.1 In fact, infancy has become the locus of much theorising in the light of findings suggesting that, shortly after birth, children display much complex perceptual and cognitive knowledge of the physical world surrounding them, which was not at all predicted to exist so early in the child’s life by previous theories. These precocious capacities include the perceptual discrimination and the categorisation of entities in a number of domains, the discrimination of spatially relevant relations, the concept of object permanence, notions of gravity, of causality, of temporality, and so on. The discovery of such capacities has led to two types of interpretations. According to the first interpretation, such complex knowledge structures are innately available and used immediately after birth. According to the second interpretation, such knowledge structures are the result of an active process of perceptual activity, which takes place very quickly after birth and allows the child to apprehend and construct the world.
2.1.1.3 Mechanisms
The extent to which innateness is or is not postulated to account for language acquisition in the form of pre-programmed equipment has consequences for the types of mechanisms that are assumed to underlie developmental change in child language after birth. The nativistic position postulates either no learning pro- cess at all or a relatively limited set of deductive processes that are assumed to be necessary and sufficient to account for the discovery of (innately available) structural properties of language. The most extreme nativistic positions view changes in child language as directly reflecting a process of physiological maturation. Less extreme views postulate that, given their biological endowment, which provides children with all universal properties of linguistic structure, children must nonetheless make some inferences on the basis of the linguistic input that surrounds them, partic- ularly in order to discover the particular properties that characterise their native language.
In sharp contrast, the behaviourist position, at least in its (now somewhat out- dated) extreme form, postulates that children must construct all of language, since they are not endowed with any equipment whatsoever at birth. This construction
Free ebooks ==> www.Ebook777.com
Some issues in theories of language acquisition process is seen as involving gradually more complex associations between stimuli and responses as the child interacts with its environment. Different constructivistic views postulate some innate ‘predispositions’ on the part of the infant to learn the rules of language and/or of other behaviours in other domains. These intermediary positions vary with respect to the learning mechanisms they postulate, partly as a function of the type of initial equipment they attribute to the infant. Cognitivist models see language development as the consequence of a more general cognitive development, which is itself determined by biologically pre-programmed endoge- nous processes and underlies all of ontogenetic development across domains of knowledge. Interactionist models see social interaction as the most important source of learning mechanisms underlying all of child development, including language, cognitive, and/or social development. According to this position, developmental processes stem from an external social source, resulting from a process whereby interpersonal patterns are internalised.
2.1.2 Criterial properties of language
The second important debate among theories of language acquisition concerns the fundamental nature of human language. As noted in Chapter 1, the- ories diverge with respect to which specific properties of human language should be considered as criterial in differentiating our linguistic system from the com- munication systems of other species. For the sake of presentation, I briefly and somewhat brutally sketch a contrast between structural and functional approaches to language, leaving aside details for a later discussion below (Section 2.2 below, as well as Chapter 3).
2.1.2.1 Structural approaches
Structural approaches, best represented by Chomsky’s (1981) theory of Universal Grammar, focus on the syntactic structure of language. The general aim of these approaches is to account for our linguistic competence in terms of general and specific properties of grammar. In its original formulation, the aim of Chomsky’s theory (1957, 1965, 1968) was to define a finite system of components and rules that could generate an infinite set of grammatical sentences (and no ungrammatical ones). Central components of our linguistic knowledge included the capacity to judge grammaticality, to establish relations of equivalence between synonymous sentences, or to apply grammatical rules recursively. From a developmental point of view, the logical problem oflearnabilityhas been at the centre of this type of approach. Roughly, the theory puts forward innateness as a solution to this devel- opmental problem, which can be formulated as follows (also see Saleemi 1992):
how can we account for children’s grammatical competence, given that it seems to
www.Ebook777.com
Theoretical issues
be manifest quite early, that children cannot learn an infinite corpus of grammatical sentences, and that they are not provided with sufficient evidence in the input that would enable them to determine the properties of linguistic structure? Available evidence that might help children discover their language could be of two types, depending on whether it provides them withpositiveornegativefeedback. Typical examples of positive evidence are corrections provided by adults when children make errors, showing them what the correct grammatical forms should be. How- ever, other forms of positive evidence have been proposed, including expansions of incomplete segments or other types of reformulations providing surrounding struc- tures (Gallaway and Richards 1994; Harris 1992). Although such adult interventions have been shown to occur, structuralists argue that they are neither necessary for lan- guage acquisition nor frequent and systematic enough to provide sufficient evidence for a successful learning mechanism. Negative evidence could take the form of un- grammatical sentences, providing children with examples of what is not allowed by the grammar. This type of evidence is at best rare in the input heard by the child.
Furthermore, even if it were frequent, it could not constitute sufficient evidence for relevant linguistic properties, since it would not provide the child with the criterial information concerning the precise nature of the error, except if it took the form of metalinguistic statements (hardly likely to be efficient at the youngest age). This general question has led to several versions of innate theories in which knowledge of the basic grammatical structure of human language is present at birth. In such theories, and despite variations across them, this type of knowledge is presumed to be universal and species-specific, notwithstanding some cross-linguistic variations accounted for in other ways, particularly by means ofparameters(see Chapter 3).
2.1.2.2 Functional approaches
Functional approaches conceive of language as a tool for communica- tion. The study of language, then, consists of determining how linguistic devices serve as means for particular types of goals in particular types of contexts. Two properties are fundamental to all functional approaches, notwithstanding large vari- ations across them: multifunctionality and context-dependence. Multifunctionality is linked to the fact that language is seen as having two broad functions, a repre- sentational function and a communicative function, which make it into a power- ful semiotic system, simultaneously serving to represent propositional information and to regulate social interaction. Furthermore, language is intrinsically linked to its context of use. In this respect, the structure of all languages directly encodes context-dependence by providing devices that anchor speech to its context of ut- terances, indicating the personal, temporal, and spatial parameters of the speech situation. Clearly, the focus on these properties implies that the particular linguistic
Some issues in theories of language acquisition phenomena that are of most interest to functionalists only partially overlap with those that are of most interest to structuralists. Functionalists are concerned with numerous and varied phenomena, some of which are considered by structuralists to be simply outside of the scope of ‘grammar’ per se and therefore outside of the scope of linguistic theory. One aspect of the debate is therefore to determine what should or should not be of concern in a theory that aims at accounting for the universal properties that are specific to human language. Another aspect is to confront these two approaches with respect to a subset of grammatical phenomena that are at the core of both of them (although for different reasons). Examples will be provided below when we discuss different approaches to phenomena such as pronominalisation, spatial semantics, or temporal-aspectual markings.
2.1.3 Structure vs. process, competence vs. performance
Tightly related to this differential focus on various properties of lan- guage is the relative importance attributed by different theories to two further aspects of language and of language behaviour. First, some mainly focus on the properties of linguistic structure, as these properties might be established at a given point in time, while others are concerned with the processes of linguistic change whereby this structure comes about historically or during ontogenesis. Second, theories vary in whether they focus on speakers’ linguisticcompetence, which is characterised in terms of their abstract knowledge of grammar, vs. a variety ofperformancefactors that might influence their use of language in particular contexts.
2.1.3.1 Structure and process
In structural approaches, language is seen as a global and self-sufficient structure of interrelated elements and rules. The system-internal properties of this structure are assessed from a synchronic perspective, which examines language (language-in-general or particular languages) as it stands at a given point in time.
With few exceptions (e.g. Lightfoot 1991), this type of analysis is typically carried out independently of the previous or future states exhibited by the linguistic sys- tem, as well as independently of factors that exist outside of the system and that might determine its nature at the particular point chosen for analysis. In contrast, diachronic analyses focus on language change and on the reasons for this change, invoking both system-internal properties of language and system-external (cogni- tive, socio-cultural) factors that might exert pressure on the system and thereby motivate change. In this second view, language is seen as a dynamic entity, which is always changing through time, and it is the process of change that explains the very nature of the system.
Theoretical issues
These different foci can be observed if we consider how different theories address questions related to language acquisition. As noted above, most nativistic models place heavy emphasis on the grammatical structure of language, which is viewed as being biologically pre-programmed in the child. This characterisation applies espe- cially to the most extreme nativistic models, which postulate that language acquisi- tion is entirely determined by physiological maturation. However, it also underlies other less extreme nativistic views, even though they make some room for develop- mental mechanisms, such as additional inferences allowing the child to discover the structure of language. Yet other proposals (Pinker 1979, 1984, 1989) focus on the structure of child language, although some of these models differ from nativistic models in two major ways. First, they do not define this structure exclusively in syntactic terms, but rather as a bundle of intricately related semantic and syntactic properties. Second, some do not view this structure as being pre-programmed from the very beginning, but rather as resulting from a process of construction on the part of the child, who is seen as being ‘bootstrapped’ by syntactico-semantic language properties during the course of development. Finally, a third type of model, which is more functionally inclined, focuses on the process of language development, in- voking a variety of factors and types of knowledge to account for the course of this development. Among these factors are numerous socio-cultural and/or discourse pragmatic factors, not considered to be part of the system in structural approaches, and viewed by functional approaches as centrally determining how child language emerges and evolves (Bates 1976; Bates and MacWhinney 1982; MacWhinney and Bates 1989; Karmiloff-Smith 1979, 1985, 1987, 1992). This third view is clearly process-oriented and it further differs from the other views described above in that it is centrally concerned (rather than at best peripherally) with factors determining how language is used in relation to context.
2.1.3.2 Competence and performance
Since Chomsky’s (1957, 1965, 1968) original formulation of Universal Grammar, studies of language or language development have often made a distinc- tion between two aspects of speakers’ linguistic behaviour: theircompetenceand theirperformance. Grammatical knowledge constitutes the stuff of their linguistic competence, which is brought to bear on any occasion of language use. Such a view applies to a number of models, regardless of whether they view this competence as being available at birth or constructed by the child over time. Performance factors which might influence language use are all considered to be external to this basic competence, such as various difficulties linked to young children’s immature cogni- tive system or pragmatic factors linked to the functions of linguistic devices in par- ticular contexts of use. In contrast to such a view, functional developmental models