Glauser (1985)

Glauser acknowledged that the analysis of semantic fields in data from dialect surveys is problematic because ‘the members of a semantic field cannot be collected exhaustively unless they are known beforehand’ (Reference Glauser1985: 37). Nevertheless, he noted a number of patterns in the distribution of the words lad and lass in responses to SED questionnaire items designed to elicit words for BOY/ GIRL and SON/DAUGHTER. Glauser noted a clear isogloss reflecting the ‘North-South’ divide, whereby lad was reported almost exclusively north of a line running from southern Shropshire to the Wash. South of this line, the main variant was boy, whilst kid and chap appeared sporadically. When words for SONS were elicited, the variant lads was only found in the north, but in this case sons was also found throughout the country.

A more complex picture emerged with regard to lass, since there were more variants elicited for GIRL than for BOY: girl, lass, wench, maiden, maid, mawr, mawther, chiel and missy, but lass was still mainly confined to the North and predominant in that region. This distribution seems to have been fairly stable since the early nineteenth century, as a mapping by Maguire (forthcoming) of data from Alexander Ellis's The Existing Phonology of English Dialects (Reference Ellis1889) shows a very similar pattern. Maguire points out that since Ellis's informants, like those of the SED, tended to be elderly, the language recorded in his 1889 study is that of the earlier part of the nineteenth century, and that here, as in the SED data ‘lass is typical of the northern English counties’ (Maguire, forthcoming).

With regard to DAUGHTER, Glauser found lasses confined to the north, but even here outnumbered by instances of daughters.

Overall, Glauser noted a wider currency for lad than lass, with combinations of lad/girl much more widespread than those of lass/boy. He suggested that this implied a pattern of change in progress, whereby use of lass becomes variable before that of lad and predicted that lass would be lost before lad. With regard to the geographical patterns that emerged from his study, Glauser concluded that ‘the boundaries in question are the result of social, systematical, variational and geographical forces’ and suggested that ‘we should now check with linguistic reality thirty years later’ (Reference Glauser1985: 55).

Given the reluctance of variationist sociolinguists to engage with lexical variation, it is perhaps not surprising that Glauser's challenge has hitherto not been met. The prevalent variationist research tools of semi-structured interview, reading passage and word list were designed primarily to elicit phonetic variation, informing the researcher about the pronunciation of words rather than the choice of one lexical item over another. However, as we shall see in the following sections, recent innovations in sociolinguistic methodology have made it easier to find comparable lexical data from speakers of varying age, gender, social class, region etc.

In the following sections, we present the results of analyses of these data sets, addressing the following research questions: whether lad is more widely used than lass; what patterns of variation can be observed in the data between the use of lad/lass and synonyms such as boy/girl; and whether this variation correlates with age, gender and/or social class. We examine the ‘linguistic reality’ of data collected in the late twentieth and early twenty-first centuries to see whether Glauser's predictions concerning variation and change in the use of lad/lass have been realised.

Lads and lasses in Newcastle

Whereas the SED data analysed by Glauser (Reference Glauser1985) had been elicited by means of a structured questionnaire, the data from Newcastle discussed in this section was collected in 1994 for the Phonological Variation and Change (PVC) project, directed by Lesley Milroy at Newcastle University (Milroy et al., Reference Milroy, Milroy and Docherty1997). All the participants recorded for this project (35 in all) had been born and raised in the city of Newcastle upon Tyne and recruited using the ‘friend of a friend’ method (Milroy and Gordon, Reference Milroy and Gordon2003: 32) with the aim of constructing a balanced sample according to age, gender and social class. The methodology used, which involved dyads in unstructured conversation being recorded by an unobtrusive researcher, yielded a large number of tokens of words for BOY/GIRL; SON/DAUGHTER, because these dyads tended to talk about friends and family. This data was not orthographically transcribed in its entirety by the PVC researchers, because, as the title of the project indicates, they were interested in analysing tokens of specific phonological variants. However, transcribed texts of the conversations recorded for the PVC project have since been made available as part of the Diachronic Electronic Corpus of Tyneside English (DECTE), which has enabled us to search for occurrences of lexical items in context. The PVC data transcribed for the DECTE corpus constitutes 208,295 words from 17 hours, 34 minutes and 25 seconds of recording in total (http://research.ncl.ac.uk/decte/pvc.htm accessed 31/07/12).

In the case of the data from the PVC project, a simple count of the tokens of lad/lass used by each speaker would not be meaningful, because these would not occur at a consistent rate in free conversation. Instead, we have compared the number of tokens of lad/lass with those of their Standard English synonyms boy/girl. Figure 1 shows the proportion of lad vs. boy and lass vs. girl used by speakers grouped according to age and gender.Footnote ⁴

Figure 1.

Use of lad/boy and lass/girl according to age and gender (from Beal et al., Reference Beal, Burbano-Elizondo and Llamas2012: 86)

We can see here that every group except the middle-aged males uses lad more than lass, but that lass has not fallen out of use. Moreover, the group with the highest use of lass is the young males, so it would appear that, contrary to Glauser's prediction, use of lass is not declining. So, with regard to the Newcastle data, our answer to the first research question is yes, lad is at least proportionally more widespread than lass. There is no clear pattern according to age, except that the middle-aged group use both words less, possibly a classic example of age-grading, whereby those most involved in the job market tend to use the fewest vernacular variants. With the exception of the middle-aged speakers, males use both lad and lass more than females of the same age group do, so there appears to be variation according to gender. What complicates the picture here, compared with that from the SED data, is that both lad and lass can be used with a much wider range of referents than just BOY/GIRL and SON/DAUGHTER, so they are used in variation with a number of other words.

Figure 2 shows all instances of the variants lad and boy in the major senses in which they co-vary; meaning SON, YOUNG MAN and SEXUAL PARTNER. In the case of SON, both also co-vary with the word son itself, and in the case of SEXUAL PARTNER, the words boyfriend and husband are also used. Here we can see that lad for SON is only used once; for PARTNER it is more common than boyfriend, but less common than husband. However, the vast majority of uses of lad are with the meaning YOUNG MAN. Where older males are referred to, man and bloke are the most common words.

Figure 2.

Use of boy and lad in various senses by PVC speakers (after Beal et al., Reference Beal, Burbano-Elizondo and Llamas2012: 87)

Examples of the word lad used in these different contexts are:

1. 1. The only, the young lad at the time, their only son, I used to have to look after him. (=SON)

2. 2. I-mean most of the lads worked in Shipyards, or in the pits (=YOUNG MAN)

3. 3. She went, she went with this lad when she was thirteen (=SEXUAL PARTNER)

In the first example here, lad may not be being used as a direct synonym of son, but it has the same referent as the word son in the same sentence. This is the only example of lad for SON in the whole sample. The third example could be interpreted as ‘YOUNG MAN’, but we have included it in the category ‘SEXUAL PARTNER’ because of the clear sexual connotations: ‘went with’ here is a euphemism for ‘had sex with’. Especially when used by younger speakers, the words lad and lass often refer to potential if not actual sexual partners.

Figure 3 shows the distribution of instances of lass and girl across semantic categories. The distribution is similar to that of lad, with YOUNG WOMAN being the most frequent meaning. In this case, though, girl is a much stronger rival than boy was for lad. Lass for DAUGHTER is as rare as lad for SON, and the proportion of lass to girlfriend for SEXUAL PARTNER is similar to that of lad to boyfriend. However, lass is also used for WIFE, whereas lad for HUSBAND is rare.

Figure 3.

Distribution of lass and girl across semantic categories (after Beal et al., Reference Beal, Burbano-Elizondo and Llamas2012: 87)

Examples of lass with these different meanings are:

1. 4. I mean, when I tell our lasses now … ‘Oh mother, that was the olden days' I says ‘Aye, well you should have had some of it!’ (=DAUGHTER)

2. 5. I-mean, when, even when there was lasses in my college, I never got put with any of them (=YOUNG WOMAN)

3. 6. Wor lass was drinking Purple-Passions all night (=SEXUAL PARTNER)

Figures 4 and 5 show variation between lad/lass and boy/girl by younger speakers and older speakers respectively, according to the gender and social class of speakers.

Figure 4.

Distribution of lad, lass, boy and girl according to social class and gender (younger speakers) (from Beal et al., Reference Beal, Burbano-Elizondo and Llamas2012: 88)

Amongst the younger speakers (figure 4), middle-class subjects use both lad and lass less frequently than their working and lower-middle class peers. The exceptions here are the two working-class females who never use lad or lass. This dyad is atypical in that it is made up of two young British Pakistani women, who use girl and boy exclusively to refer to their peers. In fact, the middle-class speakers use few tokens of any of these words, which could simply be a consequence of their topics of conversation. Apart from the two British Pakistani women though, all the working-class speakers use either lad or lass. Amongst the older speakers (figure 5) it is also the case that all the working-class speakers use either lad or lass, but other than this no clear patterns emerge with regard to variation according to class.

Figure 5.

Distribution of lad, lass, boy and girl according to social class and gender (older speakers) (from Beal et al., Reference Beal, Burbano-Elizondo and Llamas2012: 88)

Apart from a greater tendency for working- and lower-middle-class speakers to use lad and lass, there seems to be no firm correlation between social class and variable use of these words. This is illustrated clearly in examples 7–15 below, all taken from the conversation of one young, male lower-middle-class dyad, one of whom consistently refers to his partner as wor lass (7–12), whilst the other uses the/my girlfriend (13–15). These two young men are friends, matched in terms of age and social class. This suggests that use of lad/lass is just as likely to be a matter of style or stance as one of sociolect, and that a qualitative, discourse analysis approach to lexical variation might shed more light on this.

1. 7. Wor lass gans ‘Oh, you're not going to drink too much of that are you?’

2. 8. they're probably the canniest people you'll ever meet coz you know fine well they're not going to <XX>Footnote ⁵ your lass

3. 9. I know. Like wor lass wears a ring on that finger

4. 10. wor lass was drinking Purple-Passions all night.

5. 11. I found out yesterday that my mate's lass is up the <XX>

6. 12. wor lass gans ‘you live in a fairy tale world’

7. 13. Got to go over to the girlfriend's concert first though.

8. 14. I don't even know .. my girlfriend's birthday.

9. 15. Bet it's my girlfriend

Lad and lass in the Sunderland data

The lexical data from Sunderland analysed by Beal and Burbano-Elizondo (Reference Beal and Burbano-Elizondo2010) was collected by Burbano-Elizondo (Reference Burbano-Elizondo2008) for her study of language and identity in Sunderland. The population sample consisted of 39 speakers stratified according to age and gender.Footnote ⁶ The speakers were divided into three age groups: 16–30, 31–50 and 50+. The data was collected using the Survey of Regional English (SuRE) methodology, originally designed by Llamas (Reference Llamas1999) in order to facilitate the collection of ‘informal speech from the informant (from which multi-levelled analyses of both regionally and socially comparable data are possible)’ (Reference Llamas1999: 97). The primary tool of this methodology, the Sense Relation Networks (SRNs), enabled the systematic elicitation of lexical data from related semantic fields and has since been successfully adapted by the BBC Voices project as described by Robinson (this volume). The SRNs (see figure 6) are designed on the assumption that words are interconnected in speakers' minds, forming a ‘web of words’ (Aitchison, Reference Aitchison1994: 82–98). In this conceptual web, words are interrelated by means of semantic links such as co-ordination, collocation, superordination and synonymy. Such links, some stronger than others, organise words into semantic fields which, in turn, are also interconnected. Following this model, in the SRNs the notions or concepts for which lexical variants are to be elicited are grouped by semantic fields, and these, in turn, are related to a broader semantic field which gives a title to each SRN.

Figure 6.

‘People’ SRN

Unlike the SED questionnaire where informants were read each question and expected to provide the first word they could think of, in the SuRE methodology informants are given about one week to complete the SRNs with as many words they use as they can think of and they then discuss these in an interview with the researcher, who is able to ascertain who would use the different variants and in what contexts. Llamas (Reference Llamas1999: 102) argues that grouping notions by subject matter (i) favours the production of spontaneous responses (Johnston, Reference Johnston, Kirk, Stewart and Widdowson1985: 85), (ii) avoids any possible ambiguity and (iii) defines and delimits the meaning of each word given within each network. This is because by asking informants to think about their variants for a particular concept within a particular semantic network, they are encouraged to think about relationships between words and relationships within and between semantic fields, e.g. by thinking about words they use for ‘child (boy or girl)’, they might produce the variant kid (an affective synonym of child) alongside the hyponym lad and its co-hyponym lass, as well as the Standard English synonyms for the latter two, boy and girl respectively. In addition to this, by thinking about all of these variants as part of a ‘relations and relationships’ network, they might also produce them as variants of the related notion ‘brother/sister’ (see SRN in figure 6). Thus, whilst the Sunderland lexical data are comparable to Glauser's (Reference Glauser1985) SED data in that they both consist of speakers' responses to a questionnaire, the SuRE questionnaire might be more effective than the SED in providing some insight into the ‘semantic structure of the words elicited’ and the members of particular semantic fields (e.g. speakers' collocations of words), a flaw Glauser (Reference Glauser1985: 37) identified in questionnaires employed in classical dialect surveys. On the other hand, however, the Sunderland data differed from the Newcastle PVC data in that the latter consisted of lexical items recorded in free conversation. Whilst the methodology used by the PVC project has the advantage of providing tokens of the words in context, the SuRE methodology provides a greater amount and range of tokens, as the terms for the relevant concepts are explicitly elicited, and, as we shall see, the methodology encourages the elicitation of related words such as lad and lass.

The Sunderland lexical data analysed in this section was elicited from the ‘relations and relationships’ network in the PEOPLE SRN (figure 6). Lad and lass and a range of synonyms were recorded as variants for the notions: ‘partner (sexual) male/female’, ‘child (boy/girl)’ and ‘brother/sister’. Nevertheless, in the Sunderland interviews, some of the informants also provided evidence of lad and lass used to mean ‘young woman/young man’, a meaning which was also present in the PVC data analysed above. Some examples included:

1. 16. I went to school with a lad called ‘lanky Robinson’ (OM10 - 28:15)

2. 17. […] and then you tend to think, well, y- you're talking to the lads in the pub or work […] (OM10 - 29:00)

3. 18. This lad was dancing with this lass… (OM27 - 23:10)

Lad and lass were frequently elicited as variants of ‘child (male/female)’. As can be observed in figures 7 and 8 facing, these were the only gender-specific responses provided by the male informants. However, both terms were less frequently elicited from the females who, instead, also produced other gender-specific variants such as boy and son for a male child, and girl and daughter for a female child. By far the most frequently elicited words which both genders entered in the SRNs for this variable were gender-neutral terms such as kid(s), bairn(s), young one and toddler. It was interesting to note that the parallelism between lad and lass in both gender groups (each variant was elicited eight times by the males and three by the females) was the result of eleven speakers (eight males and three females) producing both terms as a semantically-related pair (antonyms). This also applied to boy/girl and son/daughter, supporting the argument that the SuRE methodology encourages informants to think of words which are related to one another.

Figure 7.

Distribution of variants of ‘child (male)’ in the Sunderland data by gender

Figure 8.

Distribution of variants of ‘child (female)’ in the Sunderland data by gender

As shown in figures 9 and 10, the responses for ‘sexual partner’, either male or female, showed that lad and lass compete with a larger number of variants: boyfriend/girlfriend, husband/wife, the other half, old man, missus, dad/mam. Interestingly, the male participants were not as forthcoming at providing their words for ‘male sexual partner’ as the females were with ‘female sexual partner’. Whereas, for the former variable, males produced a total of 10 responses and the females 33, for ‘female sexual partner’ males produced a total of 30 responses and females 26. This suggests that perhaps the male participants were not as comfortable providing words for ‘male sexual partner’. However, the proportional number of elicitations for each variant was generally similar for males and females, especially in the data elicited for ‘female sexual partner’.

Figure 9.

Distribution of variants of ‘male sexual partner’ in the Sunderland data by gender

Figure 10.

Distribution of variants of ‘female sexual partner’ in the Sunderland data by gender

For the female informants, the two most frequent responses for ‘male sexual partner’ were husband and boyfriend, with ten and eight tokens respectively, followed by lad and (old) man (five responses each), other half (four occurrences) and dad (one occurrence). The males' responses, by contrast, did not show a great deal of variation, as no variant displayed more than three occurrences. In this group, husband and (old) man were the most frequent words with three tokens each, followed by lad, boyfriend and other half (two, one and one occurrences respectively). In all male and female responses, lad was always preceded by either my or our.

Lass was the most frequently elicited variant from both males and females, although females entered this response in the SRN more frequently than males. With one exception, lass was always preceded by our, just as with lad. The second and third most frequent words were girlfriend and wife, with the females providing these responses seven and four times respectively, and the male informants three and eight respectively. Other half, missus and mam only occurred once or twice in each group.

Given the parallelism between lad and lass (and other lexical pairs) in words for ‘child’, and the tendency for SRNs to encourage such parallelism (see above), it was interesting to find that, just as with the Newcastle data, lass was elicited more frequently than lad in the sense of ‘sexual partner’ by both male and female subjects. This is contrary to Glauser's (Reference Glauser1985) finding that in the SED data lad was used more widely than lass. However, we must remember that his study of variation of lad and lass focused on the questions designed to elicit variants for ‘boy/girl’ and ‘son/daughter’ and not on those designed to elicit variants for ‘husband/wife’, for which lad and lass were not produced as responses. This absence of a parallelism between the male-specific and female-specific variants in the Sunderland data was also noted in all of the other words elicited for this variable. In the case of ‘sexual partner’, elicitation of one term did not always lead to the elicitation of its pair as it did when these items were variants of ‘child’. The informants' responses might have been influenced by their actual language usage, as some male participants indicated in the interview that they did not know what terms female speakers would use to refer to a male sexual partner. This is perhaps also supported by figures 11 and 12 where boyfriend and girlfriend display an indirect correlation with age, whilst husband and wife show a direct correlation with age, suggesting that speakers in each age group produced terms which reflect the type of sexual relationship to which they are more likely to be committed at their age, i.e. boyfriend/girlfriend in the younger group, and husband/wife as age of the informants increased.

Figure 11.

Distribution of variants of ‘male sexual partner’ in the Sunderland data by age and gender

Figure 12.

Distribution of variants of ‘female sexual partner’ in the Sunderland data by age and gender

The most frequently elicited words for the SRN notion ‘brother/sister’ were either brother/sister, (our) kid, our + [FIRST NAME] and our boy. Lad and lass were rather rare variants. Each of them was produced by only one female and two males, all of whom were over 50 years old, with the exception of one male in the 31–50 age group who produced one of the lass tokens.

After analysing the distribution of lad and lass in the Sunderland data, if we extract the data specific to these two variants (see table 1), we can draw two main conclusions: firstly, and in line with the Newcastle data, there appeared to be some variation according to gender as, overall, males used both lad and lass a lot more than females, especially in the sense ‘child (male/female)’; and, secondly, whilst lad and lass occurred at equal levels when meaning ‘child’ or ‘brother/sister’, lass was elicited a lot more frequently than lad in the sense ‘sexual partner’ by both males and females. This seems to suggest that lass is today more frequently used for ‘female sexual partner’ than for ‘female child’ or ‘sister’ and may have, therefore, undergone some semantic change. Whilst this trend seems to contradict Glauser's predictions that lass would disappear before lad, figure 13, by contrast, does provide us with some evidence of change over time having taken place in Sunderland as usage of both lad and lass in all the meanings analysed appears to decrease with speaker age. However, we must remember that the figures in this graph reflect the number of occurrences of each variant and therefore it is difficult to draw any solid conclusions from such small numbers.

Figure 13.

Overall distribution of lad and lass in the Sunderland data according to age and meaning

Table 1:

Occurrences of lad and lass in all three senses (‘child’, ‘sexual partner’ and ‘brother/sister’) by gender