Lecture Methods of Lexicological Research

Statistical methods
Statistical linguistics deals with the quantitative study of language
phenomena. Its results can be used for verification.
Statistical methods are applied in the analysis of different structural types of
words, affixes and the vocabularies of great writers. They also help to select the
most frequent items for teaching purposes. Statistical regularities can be observed
only if the phenomena are numerous and their occurrence frequent, e.g. it was
found that about 1,300 – 1,500 most frequent words make up 85% of the text.
Statistical methods of analysis have been also applied to various theoretical
problems of meaning, e.g. G.K. Zipf discovered that the number of the word‘s
meanings is equal to the square root of its relative frequency.
Statistical approach is quantitative, while most linguistic problems are
qualitative. We should Know what to count. Even if we just want to count the
number of lexical units, we have to define what a lexical unit is – whether we must
count the following as one or more units: work n – work v; by chance; at large,
etc.
Modern structural ways of analysis are often combined with statistical procedures. Statistics describes how things are on the average. For a modern linguist it is not enough to know that it is allowable for a given structure to appear, he is interested in its frequency, in how often it appears. Every lexicological research is based on collecting linguistic evidence, i.e. examples.
Having determined the object of research, the problem to be investigated and the set of units or phenomena to be described, the linguist proceeds to choose his method and collect and classify his data. He must have at hand a sufficiently wide choice of contexts so that his results might be statistically reliable. To know how many examples are necessary to make the conclusion, one must determine the relative frequency of the phenomenon or unit studied.
Mathematical statistics supplies the research workers with formulas showing the necessary scope of material depending on the amount of error they are prepared to tolerate.
When using a statistical method, it is true that some details are lost because statistical study is necessarily simplifying and abstract. G.Miller gives a clear picture of the situation when he says “At one time we look at the talker as generator of sound waves, and at another time he seems a fountain of prepositional phrases. The choice depends upon the interest”.
Contrastive analysis is a detailed comparison of the structure of a native and
a target language. It is the basis of teaching foreign languages: it helps to force
and prevent recurrent mistakes caused by the interference of the learner‘s mother
tongue.
Contrastive analysis reveals sameness and difference in the lexical meaning
and semantic structure of correlated words. Thus, English kinship terms do not
always differentiate between male and female: cousin – двоюродный брат,
двоюродная сестра. The semantic structures of correlated polysemantic words do
not coincide, e.g. in English the word head is used to denote the head of a person,
bed or match, whereas in Russian different words have to be used: голова
человека, изголовье кровати, сторона монеты. The number and meaning in the
correlated sets differ: pitiful – жалкий, жалостливый – pitiful. Correlated words
also differ in their valency: new and новый are correlated only in some
collocations: новое платье – a new dress; others have to be specially learnt by
Russian students: new potatoes - молодая картошка, new bread - свежий хлеб.