Language acquisition: theories



A number of theoretical positions can be identified:

· Language is a set of habits, with associations formed between words and the real-world states/objects which they refer to. See behaviorism.

· Language is acquired through extended exposure to adult speech and a desire to make sense of the environment. See empiricism.

· There is an innate language faculty, which is (a) fully developed at birth; or (b) programmed into the maturation process. See nativism.

· A general cognitive predisposition equips infants to trace patterns in the miscellaneous language which they encounter. See cognitivism.

· The cognitive limitations of the infant equip it for cracking the language code. See ‘less is more’.

· Language is acquired through the infant’s desire to interact with its carers. See social-interactionism.

The original behaviourist view is much out of favour, but the recent emergence of connectionism has raised again the possibility that language is acquired by a process of association and without the need of cognitive pattern-recognition skills.

SEMINARIV

Tasks:

1. Read the text “Selective Processing in Text Understanding”and give definitions of the terms in bold.

2. Choose any piece of text (5–6 sentences), prove the coherence of this text. What types of inferences does the text include?

3. Prepare 5 more questionson the text to check the understanding of it by your groupmates.

SELECTIVE PROCESSING IN TEXT UNDERSTANDING*

 

It is commonly assumed that the goal of reading a piece of text is to end up with a mental representation of that text. Texts which are unclear may give rise to poorly structured mental representations, while well-written texts, if read properly, will give rise to well-structured representations which are in some sense the result of the comprehension process. Whatever the details of such a process, the text should induce a coherent representation in the mind of the willing reader. Of course, the mental representation will not be a verbatim copy of the text itself unless special rote-learning procedures are adopted. Rather, subjects typically retain the gist of a text. The point is easily appreciated that one might read and remember the gist of a novel, but not generally retain it verbatim. In short, most of what is remembered is the product of comprehending the text rather than learning it.

The problem is bound up intrinsically with which inferences are drawn on the basis of a text. Ideas which were not included in a message but which are captured by the internal representation of that message are called inferences. Since a text could support any number of inferences of varying degrees of plausibility, there must be some mechanism supporting the selection of some kinds of inference over others. The simplest idea is that only inferences necessary for the establishment of cohesion are drawn, but as we shall see, this criterion leaves room for interpretation, and there is evidence that elaborative inferences are made, and that inferences necessary for cohesion are not always drawn.

Coherence in text and mind

The term COHERENCE has been used in relation to texts and to the mental representation of texts. The idea of coherence in text is closely related to the notion of well-formedness. Thus Reinhart (1980) claims that it is coherence that distinguishes a text from a set of sentences which are unconnected. If a text is seen as an object, then it is clearly important that there are visible aspects of the text which represent the connections between sentences. These connectors are termed by Halliday and Hasan cohesion devices, because they hold the text together, giving it coherence. Halliday and Hasan distinguish five kinds of tie: conjunction (e.g., and, but, because), coreference, substitutions, ellipsis, and lexical cohesion.

The presence of surface markers of this type seems to be too weak a criterion for coherence in some instances: It does not rule out sets of sentences which we would not want to consider to be coherent in an everyday sense, such as (1).

(1) I bought a Ford. The car in which President Wilson rode down the Champs-Elysee was black. Black English has been widely discussed. The discussions between the presidents ended last week. A week has seven days. etc.

The presence of coreference links (Ford–The car, black–Black, etc.) is not enough to specify what could reasonably be called a coherent text. (Enkvist,1978, calls such examples pseudo-coherent, because they do hold together in any fashion.) Reinhart (1980) suggested that the following conditions are a better specification of what might make a text coherent. The discourse should be:

1. Connected. The sentences (clauses) of a text will be formally connected, in that each adjacent pair is either referentially linked, or linked by a semantic sentence connector.

2. Consistent. Each sentence has to be logically consistent with the previous sentence.

3. Relevant. Each sentence must be relevant to an underlying discourse topic and to the context of the utterance.

The first criterion of formal connection means that cohesion markers should be present. Again, it is possible to look for these in a text itself. The second is more problematic, for while it is possible to test for inconsistency (e.g., contradictions) in a text, it is not possible to test for consistency. Presumably, the second criterion is that each successive sentence is not inconsistent. Moving to relevance, it is not clear how relevance can be found in a text. Rather, it must be the product of inference and so makes Reinhart's definition partly dependent on a reader.

We have seen that surface cohesion markers are not sufficient for a text to be called coherent. Are they necessary, as Reinhart's formulation requires?

The answer is no, because it is possible to have a text which we would want to call coherent, but which does not depend on the presence of any markers.

(2) At dinner last night, John burnt his mouth. The soup was too hot.

These utterances are connected only by way of an inference: The second sentence explains how John came to burn his mouth. Such constructions are commonplace in normal, easily comprehended writing.

In summary, coherence does not seem to be a property of text; rather it is a property of the mental representation (interpretation) of a text. A text can yield a coherent mental representation even when it does not contain appropriate cohesion markers. From now on, we shall suppose that coherence occurs in the mind of the reader and is the establishment of a mental representation which consists of a connected set of ideas based on appropriately interpreted discourse. Both connectivity and appropriate interpretation are matters of degree, and we shall examine some influence's on selective processes which underlie coherence, starting with the issue of inference control.


Дата добавления: 2018-06-01; просмотров: 279; Мы поможем в написании вашей работы!

Поделиться с друзьями:






Мы поможем в написании ваших работ!