Hudson says that quantitative studies of speech seem relevant to theoretical linguistics because they involve central areas like sounds, words-forms and constructions. He restricts the study to English (non-standard variety). It is concerned with form of words and constructions. For instance the [h] in ‘house’ and ‘hit’ is omitted by some speakers. One can say those who include it and those who omit it have different language system. Some people will say I didn’t eat any apple. Others will say I ate no apple. What can one be saying is the relation between these people’s grammar? Do other grammars differ in morphology, in syntax or in semantics? And how do one allow for those who alternate between the two. These are the questions for theoretical linguistics. Hudson uses review of ‘Labovian sociolinguistic’ or ‘the Labovian paradigm’, which says that one cannot understand how languages change unless he have an accurate view of what language systems are like (say the study of ordinary colloquial English).
This branch of sociolinguistics compares texts or people with one another. A research can be done to find out, for instance, how many people use [h] when they say ‘house’ and how many people don’t.
Why Study Speech Quantitatively?
Hudson says if each text contained instances of only one variable for each variant, then it could be located in the relevant multi-dimensional linguistic space without using qualitative methods. When (h) is investigated in a number of texts, one might find that some texts contain instances of (h):[h], but no instances of (h):ø and that the other texts contained (h):ø but no instances of (h):[h] with this, two variables will define just two clearly distinct groups of texts.
The study of a large number of texts would indicate the extent to which these two linguistic variables are sensitive to the same social variables.
Hudson outlines the problems of methodology even though it is important.
- Selective speaker circumstances and linguistic variables.
- Collecting the text.
- Identifying the linguistic variables and their variant in the text.
- Processing the figures,
- Interpreting the result.
In discussing the methodology, he considers five different works using five different methods. The first example tend to illustrate what he refers to as the personal ingenuity of William Labor (1972) who formulated an hypotheses about the use of single linguistic variable: (r), in New York. The study represents the presence or absence of (r):[r] versus [r]:ø- of a consonantal /r/ sound in words like farm and fair were the ‘text’ sound is not a vowel in the same word (as in very or red). Data collection method was suitable for the hypothesis to be tested.
Labove went to stores asking the shop-assistants where some goods were, expecting their answers to be fourth floor or on the fourth floor. He made them to say it again. He also could test the hypothesis about social status by comparing the stores with each other, since they could be ranked from high to low status based on prices of goods, status of customers, and newspaper which they advertised. None of the speakers could notice they were taking part in a linguistic research.
It was discovered that older people used the earlier variant, (r):ø the hypothesis was confirmed that the assistants used (r):[r] in both words
(1) Phonetic Variables, where the same phonological pattern has different phonetic realization.
(2) Phonological Variable, where the same items has alternative phonological structures.
(3) Morphological Variable, where the same word have alternative morphological structures (defined in items of roots and affixes).
(4) Lexical Variables were the same meaning can be expressed by two different lexical items (near synonyms).
(5). Syntactic Variable, where the same meaning is expressed by different syntactic structures. A variable is a collection of alternatives which have something in common.
Calculating scores text
The Labovian approach can be used for calculating scores but it has a fault. A score is calculated for each variable in each part of the collected data which he calls a “text”. Texts are compared with respect to one variable at a time. To calculate the text scores for a given variable, a score is assigned to each of its variants; the score for any text is then the average of all the individual scores for the variants in the text.
This method has two weaknesses when applied to variables that have more than two variants. The first is to do with ranking of variants. The second is to do with balance among variants, since the final figure for a text gives no idea of the relative contribution made by individual variation.
He suggests that figures should not be reduced to a single score for each variable, but to keep those for each variant separate as percentage of the total cases where the variables occurred.
Calculating Score for Individuals and Groups
A research might involve the study of ten variables in the speech of sixty people under four different circumstances, which would produce 10 x 60 x 4=2,400. Separate scores for text if the classical Labovian method were used. If the alternative of quoting separate scored for individual variants were adopted, it would have a much larger figure. It is a problem handling such large amount of data. One of the ways it can be done is to breaking dawn to groups average for each group. Sixty speakers may be reduced to eight groups. This will reduced the total numbers of figures from the initial 2,400 to 320.
However, he notes that the reliance on group scores conceal the amount of variation within each group.
1,609 total views, 3 views today
Recent search terms:
- the quantitative study of speech
- 41INTRODUCTION Planning (even in collecting research data) is an important part of human activities. Whatever people decide to do, for instance if a person would…
- 30Entries and exits About entering and ending an interaction, Hudson explains that speech is structured, and it is composed of grammatical structures: sentences (short or…