Many words to do with everyday life are not in general use throughout a complete language area and indeed many words of a language are known only to speakers within fairly confined regions. This realization and the related observation that common words often have· a distinct pronunciation in local speech as well as meanings which differ regionally, gave rise to the study of what words are to be found in a locality and what meanings are given to them. Of these studies the impetus to compile glossaries has been the most prevalent. These vary in coverage, some confining themselves to the uncommon words of a village or parish, and are little more than a chapter in a history of a dialect. Others aim at a complete collection of all the words in a defined locality. The value of these compilations is that they give precise information on the vocabulary used in one area at a particular time. More ambitious projects exist in some countries. In Finland, for example, collecting dialect words and phrases systematically in a number of determined locations covering the whole country in order to compile a Finnish dialect dictionary has been in progress for some years. Such glossaries and dictionaries should show with phonetic accuracy the form of a word as used by local speakers as well as defining fully its different meanings, giving examples of its usage in the various contexts where it is to be found.
The patternings exhibited in a language as spoken within speech communities which are geographically contiguous may often show imparity. The particular 'variety' of language that occurs in any community has its own structural patterning. In the phonological encoding of a language, varieties usually have a number of speech sounds in common, but in some regions sounds of distinctive quality occur which immediately identify a speaker from those areas. Furthermore, speech sounds may not all occur in the same positions in a linguistic structure in the varieties of two separate communities. Trubetzkoy distinguished these variations thus: 'A phonological difference based on inventory exists when a dialect possesses a phoneme that is not known in another dialect. A difference in phonological function is present when a phoneme in one dialect occurs in a phonological position in which it is not found in another dialect. 8
Again at another level in the language pattern, differences may occur. 'In many languages words play an important grammatical role, in that they are built out of smaller elements by certain patterns, but are put together into sentences by rather different patterns. 9 Word-forms and word arrangement in sentences in the variations of two areas showing incongruences exhibit, therefore, dissimilar morphological and syntactical patterning.
Synchronic descriptive studies aim to give an account of the linguistic encoding for a particular region. Within a chosen area homogeneous samples of speech are identified and the relationship patterns exhibited by them are analysed and described. The uniformity of the samples often determines the size of the area chosen for investigation. Conversely the study may describe the range of variation in linguistic patterning in one area taking into account the sociocultural factors that have conditioned speech and the variation related to age groups within communities. Such descriptive studies may deal with the data comprehensively or confirm themselves to selected aspects. As Leonard Bloomfield remarked ‘The modern demand would be rather for a description such as one might make of any language: phonology, syntax, and morphology, together with copious texts'.10
Whereas the descriptivist is concerned with investigating the linguistic patterning within one defined area, the comparativist sets out to gain an overall view of the linguistic variation obtaining within the whole speech area. The practices of these two students of linguistic variation contrast sharply. The descriptivist uses all possible data and very many informants within his chosen area of investigation. The comparativist uses the method of selective sampling:
'By choosing a limited body of linguistic items for investigation in a limited number of carefully selected communities, each of them represented by a single speaker belonging to a certain social class, or by one for each of two or more social levels, the area linguist hopes to obtain a general view of the dialectal structure of the total area within a relatively short time. He has no illusion about achieving a complete coverage of usage.'11
Linguistic sampling of this kind depends for its reliability on co-ordinating several factors:
The direct recording of the spoken language in each area being investigated must be undertaken by trained fieldworkers.
The linguistic items which are likely to reveal the whole spectrum of variation within the language must be carefully determined in advance. This means the preparation of a fieldwork questionnaire to be used in all the locations chosen for the investigation.
The communities within the speech area from which samples are taken must be selected so that they are likely to produce comparable data and adequate coverage.
Within each community the informants chosen to provide the local form of speech must be assessed so that they match as to social background and style of speaking.
The collecting work must be completed within a reasonable period of time so that the data reflects the state of the language at a given period.
Obviously the inclusion of more linguistic items in the fieldwork questionnaire might reveal some variations more clearly - incidental material observed in the course of collecting is often noted down - and the addition of more points of enquiry might show up in greater detail definite trends in particular areas, but the comparativist has a set purpose and a planned programme to adhere to if he is to acquire suitable data from all the communities included in the investigation. It must be stressed that this is sampling using a wide network. Any problem areas can be the subject of further fieldwork.
Collecting in the field will result in complete sets of local speech forms for each location visited. As a corollary each individual linguistic item in the survey, if the field collection has been successful, will have a local response corresponding to the location where it was noted down. These local forms can, therefore, be set out on a map of the whole speech area and this will reveal what areas use the same forms, what areas are different and what areas use more than one form. This is the arrangement employed by the comparativist and, thus, linguistic geography -or dialect geography as it is also called -came into being, resulting in complete dialect atlases covering many speech forms for a whole country.
Within a speech area, locations where the same linguistic form is used can be separated from others which use a different form. Variant phonological realizations of the same 'word' and variant 'words' having the same lexical meaning exhibit this grouping and separateness. Consequently dividing lines can be drawn on a map to mark off areas which resemble each other in their use of a particular speech form from other areas. These lines are called 'isoglosses'. Where the isoglosses for many linguistic items co-occur almost entirely across speech areas a 'dialect boundary' is said to exist along those locations. Comparative studies reveal the range of variant forms for each item within a speech area and, by examining the general pattern of many items, aim at determining where speech boundaries exist. These are then studied in relation to physical terrain, settlement patterns, development of transportation systems, growth of regional centres, etc. in order to see whether their presence across areas can be explained by such factors.
The central problem in studying areal patternings has always been the ordering of the particular features of a sample of speech so as to establish how they relate to the same particular features of another homogeneous sample of speech. The common practice was to focus on the immediately observable similarities and differences of speech forms -phonological variants of the same 'word' or lexical items - and on comparing those variant forms directly with the aim of revealing similarities or differences in the linguistic patterning. Uriel Weinreich in an important paper voiced a primary objection to such an approach. Such comparisons, he maintained, 'ignored the structures of the constituent varieties … existing dialectology usually compares elements belonging to different systems without sufficiently stressing their intimate membership in those systems'. He insisted that 'the forms of the constituent systems be understood first and foremost in terms of .those systems, 'and that structural dialectology's 'special concern is the study of partial similarities and differences between systems and of the structural consequences thereof.'12 He proposed a particular comparative co-ordering by means of what he called a diasystem. W.G. Moulton attempted to take areal dialectology a step further by mapping differences in complete systems, for example the vowel systems of different regional varieties of the language. 13