И. В. Арнольд лексикология современного английского языка Издание

§ 11.6 DIFFERENT TYPES OF NON-SEMANTIC GROUPING
The simplest, most obvious non-semantic grouping, extensively used in all branches of applied linguistics is the alphabetical organisation of written words, as represented in most dictionaries. It is of great practical value as the simplest and the most universal way of facilitating the search for the necessary word. Even in dictionaries arranged on some other principles (in “Roget’s International Thesaurus", for example) we have an alphabetical index for the reader to refer to before searching the various categories. The theoretical value of alphabetical grouping is almost null, because no other property of the word can be predicted from the letter or letters the word begins with. We cannot infer anything about the word if the only thing we know is that it begins with a p. Only in exceptional cases some additional information can be obtained on a different, viz. the etymological, level. For instance, words beginning with a w are mostly native, and those beginning with a ph borrowed from Greek. But such cases are few and far between.
238
The rhyming, i.e. inverse, dictionary presents a similar non-semantic grouping of isolated written words differing from the first in that the sound is also taken into consideration and in that the grouping is done the other way round and the words are arranged according to the similarity of their ends. The practical value of this type is much more limited. These dictionaries are intended for poets. They may be also used, if but rarely, by teachers, when making up lists of words with similar suffixes.
A third type of non-semantic grouping of written words is based on their length, i.e. the number of letters they contain. This type, worked out with some additional details, may prove useful for communication engineering, for automatic reading of messages and correction of mistakes. It may prove useful for linguistic theory as well, although chiefly in its modified form, with length measured not in the number of letters but in the number of syllables. Important statistical correlations have been found to exist between the number of syllables, the frequency, the number of meanings and the stylistic characteristics a word possesses. The shorter words occur more frequently and accumulate a greater number of meanings.
Finally, a very important type of non-semantic grouping for isolated lexical units is based on a statistical analysis of their frequency. Frequency counts carried out for practical purposes of lexicography, language teaching and shorthand enable the lexicographer to attach to each word a number showing its importance and range of occurrence. Large figures are, of course, needed to bring out any inherent regularities, and these regularities are, naturally, statistical, not rigid. But even with these limitations the figures are fairly reliable and show important correlations between quantitative and qualitative characteristics of lexical units, the most frequent words being polysemantic and stylistically neutral.
variants of these vocabularies have received the derogatory names of officialese and journalese. Their chief drawback is their triteness: both are given to cliches.