4 Activity‐based corpus analysis and design of Phrases 1
4.1 The use of the Gothenburg Spoken Language Corpus (GSLC)
70
4 Activity-based corpus analysis and design of
71
4.1.1 Vocabularies for VOCAs
The main idea behind the vocabulary described in this thesis is to provide users with frequently occurring phrases for specific activities. They are meant to be used as a complement to other methods like writing and using word prediction, and to enable users to be both fast and accurate communicators. The phases have to be organised so that they are easy to find and there is a limit as to how many phrases can be remembered or recalled by a specific user. A vocabulary should also include features that make it possible for the user to create new, unique utterances (Todman & Alm, 2003). Since the potential users form a heterogeneous group, it is important that it is possible to customise the vocabulary for each user, including the user interface that ought to be possible to change, depending on the preferences and abilities of the specific users (motor functions, perception, roles, etc.). The expressions that are included in the vocabulary described towards the end of this chapter, should be seen as suggestions that are meant to be changed by the end users, should they prefer to express themselves in some other way.
In order to know what phrases to include, one priority has been to learn more about typical conversations in the chosen activities. The activity that was chosen as a starting point was shopping: more specifically, conversations between customers and shop assistants. The main method for studying the expressions in shop conversations was to use transcriptions from the activity ‘Shop’ in GSLC, specifically the communicative acts found in the sub-activities Food and Games, but also frequencies from the whole shop activity and collocations from the whole corpus.
4.1.2 The use of communicative acts
The concept of communicative acts has been used as a unit of analysis, because it is viewed to be of use in the process of extracting expressions of many different kinds for the vocabulary. That we perform actions through our speech, speech acts, was originally proposed by Austin (1962) and later elaborated by Searle (1976) and many other sociologists as well as linguists. The concept has been very successful, but not uncontroversial. Following Austin, speech acts are often discussed in terms of their relation to the concepts of illocutionary acts/force (intentions of the speaker) and perlocutionary acts (effects on the hearer). Taxonomies of speech acts have been created and criticised, especially attempts to find the functions of the speech acts only in the grammatical structures, separated from the context. In order not to exclude non-speech modes like gestures, eye gaze and body movements, the word communicative act is often used instead of speech act. The way communicative acts are used in this study, builds on Allwood, Nivre and Ahlsén (1993), Allwood (2000a+b) and Mey (2001, 2003), but has been adapted in order to facilitate its use for the specific purpose of creating a vocabulary for VOCAs. One of the features that are important is that the same communicative act can be expressed in many different ways; with different words and grammatical
72
structures, as well as through body movements and gestures (Allwood, 2002). It is then only a small extension of the natural occurrence to express such an act through the use of an artefact, such as a VOCA.
Table 4.1. Examples of communicative acts used in shops
Communicative act Expressive Evocative Obligative (speaker)
Obligative (listener) Question & Request for
information
$A: mer påsar / va{r} ligger de{t}
more bags / where are they
$C: de{t} finns där they are there
$A: (aha)
Wonder where more bags are to be found.
That C should tell A where he can find more bags and/or provide him with one.
That A does not know where the bags are but that he wants to know and that he wants to get one.
Evaluate the question; decide if he knows the answer and if he wants to provide it.
Give A a response and/or bag.
Request for clarification
$A: tvåhundrafyrti{o}nie spänn two hundred forty ‐nine quid
$F: va{d} sa{de} du tvåhundra what did you say two hundred
$A: fyrti{o}nie forty‐nine
Wonder about what A said.
That A should provide him with the information he missed.
That F did not hear what A said and that he wants to know.
Evaluate the question and decide how much additional information F needs. Decide how to respond and tell F what he wants to know.
Evaluation, Acceptance, Acknowledgement & Ending interaction
$R: å0 så står de{t} va{d} har gjort å0 / när den var här å0 så and it says what been done and / when it was here and so
$T: jättebra / alla tiders / då tackar vi för de{t} / hej very good / wonderful / we thank you for that / bye
Expression of satisfaction, thanks and farewell.
That A should know that T is satisfied, grateful (and/or polite) and about to leave.
That T is satisfied and about to leave.
Evaluate T: s expression and decide whether or not the transaction and/or
conversation is finished. Decide how to respond.
It is important to bear in mind that the communicative acts are situated and context- dependent – they are not inherent features of the expressions. Although the same expressions are found to perform the same communicative act over and over again, the same act can be performed by many other expressions as well, and the same expression that performs one communicative act in one situation can perform a totally different act in another. Also, in a given situation, more than one communicative act can be
performed by the same expression. Other times, an expression may contain several communicative acts, expressed in succession. It is not the aim of this thesis to make any new theoretic claims regarding communicative acts; rather, it is to build on existing
73
models and use them as tools for the purpose of creating a vocabulary for VOCAs that can be of as much use as possible for its intended users.
To show how communicative acts can be studied in Allwood’s activity-based framework, some examples from the shop conversations in GSLC are given in table 4.1. In the table, the leftmost column contains quotes from the corpus, with the exemplifying utterance in bold face. The expressive dimension shows the attitude that is expressed by the
utterance, while the evocative dimension describes the attitude or action that the speaker is trying to evoke in the hearer(s). There is also an obligative dimension for the speaker that is linked to the expressive dimension and to the attitude of the speaker while performing the utterance. The obligative dimension for the receiver is related to the evocative dimension of the sender. The target expressions are displayed together with some of its context, since it is only in its context that a specific utterance can be understood.
Since communicative acts are not mutually exclusive and highly context dependant, they are not suitable to include in a fixed taxonomy, such as Searle’s (1976). Instead the four components intention, behaviour, effect and context can be used as a foundation, but there is no limit to the number of possible communicative acts (Gunnarsson, 2006). The implication is that if communicative acts are to be classified, they should target a specific activity. Despite this, Searle’s categories have been used as containers for the
communicative acts that were found in the Shop conversations in GSLC, as is described later in this chapter.