1. Background

In the year 1988, I was asked by the School of Celtic Studies, Dublin Institute of Advanced Studies, to administer a linguistic questionnaire to informants on the island of Inis Mór, the largest of the three Aran Islands, which lie in Galway Bay on the west coast of Ireland (Map 1). The linguistic questionnaire was that used for the Linguistic Atlas and Survey of Irish Dialects (LASID) in the early and mid-1950s. The two islands to the east—Inis Meáin and Inis Oírr—had been surveyed at that time, but Inis Mór had been left out of the survey. The present survey was an attempt to fill that gap. I asked two elderly individuals living near the center of the island—a brother and his sister—to serve as my informants, and we began work that summer.

Map 1. The Aran Islands, Co. Galway, Ireland. The individual townlands on the large island, Inis Mór (Árainn), are shown in their approximate geographic locations. Map by Liam Ó hAisibéil, with basemap data from Esri, HERE. By permission of Deirdre Ní Chonghaile.

Inis Mór, known to its local inhabitants as Árainn, is an island only nine miles long, with a population of some 800 people.

My initial impression, formed on the basis of information received from the local people, was that Inis Mór was divided linguistically into two halves, distinguished by perhaps half a dozen isoglosses. However, when I began my linguistic fieldwork, I was amazed at the amount of linguistic variation to be found in the idiolects of my two informants—particularly in the morphology of the irregular verbs. I felt that I needed to explore this linguistic variation with other linguistic informants from all parts of the island.

To elicit the data needed to explore the principal patterns of linguistic variation, I constructed a questionnaire containing 125 sentences in English, which were to be translated orally to Irish sentences. These sentences were designed to elicit at least three or four putative linguistic variables per sentence.

Over the next few summers, I elicited tape-recorded data from a total of nine individuals spread the length of Inis Mór and five individuals in the two smaller islands to the east. The informants were a convenience sample of relatives and friends rather than a random sample of the population on the islands. Nevertheless, I feel confident that the informants are quite representative of the population of the islands, and the biographies of the individual informants allowed me to identify external influences in their lives, such as higher education or work outside the community, which might affect the nature of their speech samples. My principal effort was to ascertain that the different age groups, both genders and all parts of the islands—especially in Inis Mór—were represented in a balanced way.

In October 2015, after 25 years of patient work, the results of the investigation were presented in the form of a 1,000-page monograph on a website of the National University of Ireland, Galway.Footnote ¹

The data displayed on the linguistic tables in the monograph present a striking picture of rich and patterned linguistic variation extending over these three small islands, with a surprising amount of variation shown in the data from the largest island—Inis Mór.

As expected, most of the isogloss boundaries were geolinguistic, though certain distributions of variant forms could be explained better by appealing to factors such as age, gender, or degree of formal education. The physical location of the birthplaces of the informants was of fundamental importance in explaining the patterns of variation, and here the physical shape of the islands and their locations relative to one another were the determining factors. The westernmost island, Inis Mór, is a long, thin island some nine miles in length, and its two eastern neighbors, Inis Meáin and Inis Oírr, each roughly circular in shape and less than half the size of the large island, continue the geographical alignment of the large island, forming a chain of islands extending to the southeast.

The population lives mainly in bailte fearainn (‘townlands’ or small clusters of houses) scattered throughout the islands. On Inis Mór, there is a chain of 15 townlands stretching the length of the island.Footnote ² On Inis Meáin, the population is located in a small cluster of townlands in the very center of the island, while on Inis Oírr, the population is located in a small cluster of townlands right on the shore itself.

While the distribution of the townlands on Inis Meáin and Inis Oírr seems to offer little opportunity for geolinguistic variation, the distribution of the townlands on Inis Mór is of vital importance for explaining the linguistic patterns.

2. Linguistic variation on Inis Mór

It was clear from the distribution of variants in the data that Inis Mór could indeed be divided into the two expected linguistic areas—a western area and an eastern area. However, each of these two halves could be split in half, so that one could finally identify a series of four linguistic areas spread the length of Inis Mór, incorporating the 15 townlands of the island.

Each of these four linguistic areas would enclose at least three or four townlands, and would provide opportunities for social interaction between neighbors, with each area enclosing at least a couple of square miles each. Furthermore, at the time when the informants were receiving their formal schooling, there were four primary school catchment areas whose boundaries matched the linguistic areas rather well. Therefore, in terms of the language acquisition of the informants, the broad neighborhoods and the schoolyards within them would offer opportunities for language acquisition among peers which would promote the development of local linguistic variants within each generation. In fact, such a conception of language acquisition in the different language areas by the different generations coming to adulthood in them explained rather well the linguistic patterns which were appearing in the data.

The households themselves could be linguistically heterogeneous, since quite often the mother of the family would herself be a stranger to that linguistic area, since she could have “married in” from a distant area on the island, from another island, or from the mainland itself. This fact of likely dialectal heterogeneity within the household further emphasized the importance of the playgroup in the formation of local linguistic variants which would provide a common local identity for the children of the area.

Before going further in our discussion of the possible social mechanisms underlying language acquisition among peers, let us examine some of the linguistic variants encountered in the data from all three islands in an effort to see just what kind of linguistic variation exists in the Aran Islands.

3. Geolinguistic variation: Phonology

Beginning with phonological variation, we find monophthongization of the two diphthongs [iə] and [uə] in the eastern half of Inis Mór in such words as siar (‘west’) and suas (‘up’), respectively, resulting in variants such as [ʃeːr] and [soːs], respectively. These are the two phonological shibboleths which islanders see as characterizing the speech of the people of the eastern half of the island. In addition, again on eastern Inis Mór, the diphthong [uə] is reduced to the neutral vowel [ə] in certain words, such as buaile (‘grazing field’), bualadh (‘beating’), and buachaill (‘boy’), but this sound change has gone unnoticed by the islanders. Its importance will be seen when we examine the lexical data.

Another noteworthy phonological variable is the backing of the low central vowel [a] to [ɑː] in monosyllabic words before certain sonorants, e.g., am (‘time’), gann (‘scarce’), thall (‘yonder’). This is a sound change which is proceeding slowly through Irish Gaelic dialects as a whole, though it has not entered the awareness of the islanders yet. Here the quality of the following sonorant determines the realization of the vowel preceding it.

Hence, in the case of the word am, speakers on all three islands eschew the sound change, producing [aːm] uniformly. In the case of the word gann, however, the speakers in the western two-thirds of Inis Mór preserve the original vowel, but speakers in the eastern third of the island, along with speakers on both islands to the east, back the vowel, producing [gɑːN]. Finally, in the case of the word thall, speakers on all three islands back the vowel, producing [hɑːL].Footnote ³

Thus we can say that the sound change has worked its way through all three islands in monosyllabic words of the phonological shape -all, has not yet begun its work in words of the shape -am, and is slowly working its way westward in words of the shape -ann.

There is another linguistic variable of which the islanders are not yet fully conscious: the affrication of the palatalized consonant [t’] in such words as teach (‘house’), tinn (‘sore’), etc., producing such variants as [t͡ʃæːχ] and [͡ʃiːN’], instead of [t’æːχ] and [t’iːN’], respectively. Superficially, its pattern of distribution in the islands resembles closely that of the monophthongization of [uə] to [oː], since it is found in easternmost Inis Mór and in both islands to the east. Nevertheless, a closer look at the data reveals that this trait is especially characteristic of female speakers, and that its distribution may have something to do with women’s speech networks.

4. Lexical variables

The data include very few lexical variables, but a single lexical variable can reveal surprising linguistic variation. For example, the term for “field” in Irish in the Aran Islands yields a very clear pattern of variation (Table 1). The word buaile is clearly dominant in Inis Mór, while the word páirc occurs in two townlands widely separated from one another. In Inis Meáin, both buaile and páirc occur, but in Inis Oírr all three informants produce the word garraí, though the male informant also produces the variant páirc.

Table 1. Lexical Variation (with morphological and phonological complexities)

*The variant páirc occurs in two widely separated townlands on Inis Mór.

**The male informant on Inis Oírr uses the variant páirc in addition.

***The plural form páirceannaí as a variant occurs in the two widely separated townlands mentioned above.

If we look more closely at the phonological representations of these lexical items, however, a further subdivision of Inis Mór is visible. All the informants in eastern Inis Mór pronounce the form buaile as [bəl’ε], with a neutral vowel in the first syllable, rather than the expected [buːl’ε] found in western Inis Mór.

If we look now at the plural forms for “field,” we will find further complications. Once again, there is an apparent east/west division in Inis Mór, with western Inis Mór opting for the more Standard Irish plural form buailte, with the plural suffix -te, while eastern Inis Mór attests a variant buailteachaí, with a triple plural suffix which combines the plural suffix -te with the plural suffixes -acha and -aí.

But there is yet another subdivision within eastern Inis Mór. The informants in the townland Corrúch produce the plural variant boilteachaí, formed on the model of the eastern singular form boile ([bəl’ε]), while all their neighbors to the east produce buailteachaí, formed on the model of the Standard Irish singular form buaile. In this case, the informants in Corrúch are being consistent, while their neighbors to the east are not.

The plural variant páirceannaí also occurs on Inis Mór, but only in the two widely separated townlands mentioned above which use the singular variant páirc.

As for plural endings on the other two islands, things are much simpler. We find both buailteachaí and páirceannaí on Inis Meáin, but Inis Oírr, however, again distinguishes itself. The male informant produces the variant páirceanna, while the two female informants produce garrantaí, the plural form of the local singular form garraí, with a triple plural suffix formed of -anna, -ta and -aí.

Thus, though at first glance the islands seem to represent a straightforward situation, with the lexical forms buaile, páirc and garraí expressing the concept “field” in English in different areas of the islands, when we include both the patterned variation in phonology and the varied plural forms, the situation is a good deal more complex. Here gender may again play a role.

5. Morphological variation

Beginning with pronominal forms (Table 2), we note the distinctive first-person plural pronoun muinn and its emphatic counterpart muinne, both blends of the pronominal forms muid and sinn. These are morphological shibboleths which islanders believe characterize the speech of the people of the eastern half of the island. In fact, the data reveal that these forms are widely spread today on the island, whatever their distribution was in earlier generations. Otherwise, the Connemara variants muid and muide are dominant on all three islands.

Table 2. Morphological Variation: Pronouns

*The variant muinn is a pronominal form based on a blend of the historically distinct variants muid and sinn.

**The variants muind and muint are further blends of the variants muid and muinn. They seem to be minor variants in Inis Mór with no apparent significance in their pattern of distribution.

***The only informants across all three islands who used the Standard Irish variant seisean were the two female informants on Inis Oírr, who each represent a younger generation with more formal education.

However, variants of a second pronominal form, the Standard Irish third person singular emphatic form seisean, reveal interesting patterns of distribution, but these non-standard Irish variants have apparently passed unnoticed by the islanders. The variant sosan is found throughout the eastern half of Inis Mór and the two other islands to the east. The variant siosan, an apparent blend of seisean and sosan, is found throughout Inis Mór, but it is not found on the other two islands.

6. Inflected prepositional pronouns

Turning now to the inflected prepositional pronouns (Table 3), we find that the third person plural pronominal forms in particular show interesting patterns of variation. Most striking are the complex, overlapping patterns of variation in the distribution of the variants for the third person plural form iontu, ‘in them.’ On Inis Mór, the variant íontab, with the long, high front vowel [iː] in the first syllable, runs the whole length of Inis Mór. In contrast, the variant iontab, with a short, lax vowel in the first syllable, begins only in easternmost Inis Mór and then runs eastward through Inis Meáin and Inis Oírr.

Table 3. Morphological Variation: Inflected Prepositional Pronouns

On Inis Mór there are also the corresponding minor variants íonta and ionta, which distinguish western Inis Mór from easternmost Inis Mór, respectively. In contrast, on Inis Oírr, we find the unique variants ónta(b) and únta, with long back vowels in the first syllable, which distinguish the older speakers from the younger speaker there, a distinction which may mark generational change.

We find similar patterns for the variants of the pronominal form tharstu, ‘past them.’ On Inis Mór, the variant thartab is both dominant and unique to the island. The variant thartu is found in easternmost Inis Mór and on Inis Oírr, though it is worth remembering that the mother of the man who produced that variant in easternmost Inis Mór was born on Inis Oírr. The variant tharab, found in easternmost Inis Mór, Inis Meáin, and Inis Oírr, is strongly, though not exclusively, characteristic of female informants, so gender may play a role in regard to this variant. Finally the youngest informant in the sample, a young woman on Inis Oírr, produces an analytic construction, thar iad, and this could be a sign of generational change.

The pronominal form rompu, ‘before them,’ produces a similar pattern of variation, though the variation seems purely geolinguistic. The variants rompab/rumpab/rúmpab are especially characteristic of central Inis Mór, and the variants rúmpú and rúmpa are characteristic of Inis Oírr. The variant rúb is unique to the middle island, Inis Meáin.

Last of all, the pronominal form acu, ‘at them,’ is unusual in that while the middle island, Inis Meáin, again stands out, it is the quality of the vowel of the initial syllable that provides the contrasting feature. The form acab in Inis Meáin has the clearly fronted vowel [æː] in the initial syllable, while its counterpart acab on both Inis Mór and Inis Oírr has the low central vowel [aː] in the corresponding syllable.

7. Comparative adjectival forms

Turning now to comparative adjectival forms (Table 4), we find a striking gradient of linguistic difference extending from westernmost Inis Mór eastward to Inis Oírr in the variants for the form níos raimhre, ‘fatter.’ Beginning in westernmost Inis Mór, we find the variant níos roímhre ([n’iːs riːv’r’ε]), with a long, high front vowel in the first syllable of the second word.

Table 4. The Phonetic Realisation of the Comparative Adjectival Construction níos raimhre ([nˈiːs rʌivˈrˈε]), ‘fatter’. (The variants are ranged according to the locality of their occurrence, beginning in the westernmost point of Inis Mór and continuing eastward to Inis Oírr.)

In the center of the island, from An Sruthán through Corrúch, we find the variant níos roimire ([n’iːs rɪm’ɪr’ɪ]) at An Sruthán and the variant níos roimre ([n’iːs rɪm’r’ɪ]) at Fearann an Choirce and at Corrúch. These variants all have short lax vowels in the first syllable and an unlenited bilabial nasal consonant immediately following the vowel. The female informant at Corrúch also produces a unique variant, [n’iːs raim’r’ε].

Moving much further to the east in Inis Mór, from the townlands Eochaill through Cill Éinne, we encounter the variants [n’iːs rɪv’r’ɪ] and [n’iːs rεv’r’ε], in which the first syllable contains a short lax vowel, but in which the following consonant is now lenited, as in westernmost Inis Mór.

On Inis Meáin, we find the same variant encountered in easternmost Inis Mór, [n’iːs rɛv’r’ɛ], but also a variant with the short lax vowel in the first syllable lowered even more: [n’iːs ræv’ɪ’r’ε].

Finally, on Inis Oírr, the two older informants produce the Standard Irish variant níos raimhre ([n’iːs raiv’r’ε]), but the youngest informant there produces a variant already encountered in easternmost Inis Mór: [n’iːs rɪv’r’ɪ]. Once again, we seem to be encountering a generational difference in speech behavior in Inis Oírr.

Bear in mind that all this intricate, patterned variation is occurring between a chain of townlands in a very limited geographical space. Inis Mór itself is scarcely nine miles long.

8. Irregular verbal paradigms

The irregular verbal paradigms in Irish Gaelic give great scope for variation, particularly because of the possibilities for suppletion and pseudo-suppletion in the verb root (Rudes, Reference Rudes1980).

The future verb form tabharfaidh (‘[X] will give’) and its negative counterpart ní thabharfaidh (‘[X] will not give’) present interesting patterns of variation. We find the positive variant [t’uːrə], with a palatalized initial consonant, running the length of Inis Mór almost to its eastern tip (Table 5a). In easternmost Inis Mór, however, the palatalized consonant is strengthened to an affricate, producing the variant [t͡ʃuːrə], and this variant extends eastward through Inis Meáin. Once again, this affrication is characteristic of easternmost Inis Mór and of Inis Meáin.

Table 5. Morphological Variation: Irregular Verb Forms

*Produced by the youngest informant (18 years of age) on all three islands.

** An informant towards the east in western Inis Mór produces the present tense form ní abraíomuinn, using the verb root abr- found widely in eastern Inis Mór.

In Inis Oírr, yet another variant is found. The older informants produce the variant [tuːrə], with the initial consonant unpalatalized, as in Standard Irish and in dialects to the south.

Thus there is one group of variants which exhibits palatalization and which extends through the two western islands, and another group, which exhibits no palatalization and is unique to Inis Oírr.

In contrast, the negative form of the same verb exhibits a sharply different pattern of variation (Table 5b). The variant [n’iː x’uːrə], with a palatalized glottal fricative for an initial consonant, is restricted to Inis Mór. The variant [n’iː huːrə], with an unpalatalized initial consonant in the verb stem, now unites both Inis Meáin and Inis Oírr. Thus Inis Meáin is linked to Inis Mór by its positive variant of the verb form, but it is linked to Inis Oírr by its negative variant.

In addition, the youngest informant on Inis Oírr produces distinct variants for both the positive and the negative verb forms, producing [tuːr’ɪ] and [n’iː huːr’ε], with a palatalized flap in root-final position in both forms. These forms are constructed analogically on the model of the imperative singular form tabhair, and not on the root of the future form tabhr-. This innovation suggests generational change.

Turning now to another verb, we find that the simple past tense variant ní dhearna, (‘[X] did not do’), found in Standard Irish, occurs only in eastern Inis Mór, where it is widespread (Table 5c). Otherwise, informants in the two western islands use the variant níor rinne/ní rinne, constructed analogically on the model of the positive form of the same verb, rinne, (‘[X] did’).

On Inis Oírr, however, the older female informant produces the variant níor dhein, ([n’iːr jεn’]), a variant characteristic of dialects to the south. In contrast, the older male informant produces the variant ní dhéan, ([n’iː jεN]), which seems partly modelled on the Standard Irish variant ní dhearna and partly on the imperative form, déan, ([d’eːN]).

The youngest informant, however, again produces two distinctive variants: níor dhín, ([n’iːr jiːn’]), and níor dhion, ([n’iːr jəN]). While the root-vowel [iː] of the first variant can be heard in variants of the root déan- found in the data for other tenses of that verbal paradigm elsewhere in the islands, the root vowel [ə] of the variant níor dhion is strikingly different acoustically from the vowel found in any other form of the root déan- in all the data from all three islands. This last variant, therefore, may signify an important generational change.

The final verb form to be examined here, the future verb form ní déarfaimid, (‘we will not say’), is from the paradigm for the very irregular verb abair, (‘say’), which offers an astonishing number of opportunities for speakers to create variants (Table 5d). However, the freedom of speakers to form a multitude of variants is constrained by possible misunderstandings by listeners of such variant tense forms.

First of all, though the Standard Irish verb form ní déarfaimid is a negative verb form, the negative particle ní causes no lenition of the initial consonant of the verb stem.

Nevertheless, every speaker on all three islands lenites the initial consonant of the verb as though it were a regular verb form. Thus, we encounter the variant ní [dh]éarfaidh muid, ([n’iː [j]eːrə mɪd’], ‘we will not say’), widely on Inis Mór and on Inis Oírr (but not on Inis Meáin). This form is acoustically quite close to the present tense variant ní [dh]eireamuid, ([n’iː [j]εr’ə mɪd’], ‘we do not say’), but distinct from the latter form both in the length and quality of the vowel of the root and in the quality of the consonant following that vowel.

However, there is a second future tense variant, ní [dh]eirfidh muid, ([n’iː [j]εr’ə mɪd’], ‘we will not say’), which competes strongly with the first variant on Inis Mór in roughly the same local areas and which also occurs on Inis Meáin. This future tense variant is homophonous with the present tense variant shown above. How can speakers who use this variant of the future verb form maintain a clear acoustic distinction between the future and the present tense forms? In fact, individual speakers manage to maintain the distinction in a variety of ways, with no confusion of verb forms.

For example, an informant in western Inis Mór maintains the distinction by using the suppletive root abr- (borrowed from the imperative forms of the same paradigm) in the present tense, thus producing ní dheirfidh muid for the future tense, but ní abraíomuinn for the present tense. The latter form is further marked unambiguously for the present tense with the stem vowel -í-.

On Inis Meáin, an informant there simply uses the present tense suffix -eann to distinguish the present tense form. Thus we have the future form ní eirfidh contrasting with the present form ní eireann.

On Inis Oírr, an informant uses the alternative root variant in the future form ní [dh]éarfaidh, a variant which reflects the Standard Irish form ní déarfaidh, but then goes on to produce the present tense form ní [dh]éarann, which has borrowed the root from the future tense, but adds the regular present tense suffix -ann.Footnote ⁴ Thus, though the root of both verb forms is homophonous, the regular present tense ending -ann removes all ambiguity as to the tense.

Finally, three informants in eastern Inis Mór solve the problem of differentiating between the two tenses by taking the opposite approach. All three use a future verb form with the suppletive root abr-, but in this case they all make a second conjugation verb of this variant, producing ní abróidh muid/muinn, a variant which contains the unambiguous future stem element -ó-. In that way, whether they treat the present tense variants as first conjugation verbs (e.g., ní abramuinn) or second conjugation verbs (e.g., ní abraíomuinn), the stem elements clearly distinguish the future verb forms from the present verb forms.

Note that in some of these examples, it is nearly impossible to tell whether the initial consonant of the verb in such constructions has been lenited (ní dhéarfaidh muid [n’iː jeːrə mɪd’]) or elided completely (ní éarfaidh muid [n’iː eːrə mɪd’]). In such cases, even the native speakers may not be completely sure of the base forms, leading to such back-formations as éarfaidh muid, ‘we will say,’ and eireann muid, ‘we say.’

With such an irregular paradigm for the verb “to say,” the possibilities for individual creativity are great indeed.

9. Linguistic variation between generations

As we have already seen, the youngest informant on Inis Oírr—an 18-year-old young woman—shows quite a few innovations in her speech. But she also produces variants much closer to Standard Irish than those attested by older informants. This reflects the educational opportunities which older informants in the sample have not had.

The younger woman had attended secondary school on the Connemara mainland and was in fact attending a university in Dublin at the time of the interview.Footnote ⁵ Thus the linguistic socialization of her generation may be far broader in its geographical scope than the linguistic experience of earlier generations on the islands or even in Connemara itself. Her own educational path is now common on the islands, with many youngsters of her generation going on to study in third-level institutions on the mainland. In this way, their exposure to Standard Irish, as well as to Standard English, is far greater than in the case of earlier generations.

Another informant in the sample, the youngest female informant on Inis Mór, also stands out from the other informants on Inis Mór. Again, it is not only her non-standard variants, including innovations as well as retentions, but her Standard Irish variants or at least near-standard variants which attract our attention. Though born on the western tip of the island, she received her secondary education at the technical school at the other end of the island, and later married into the community there. She also received post-secondary vocational education on the mainland. As a result, she exhibits less marked versions of local linguistic traits than her neighbors, variants which are closer to Standard Irish variants, and even traits of the eastern end of the island. She was 43 years old at the time of the interview, so she was roughly a generation older than the young woman on Inis Oírr.

The older female informant on Inis Oírr was roughly of the same age cohort as the youngest woman on Inis Mór just mentioned, and also received her secondary education at the same technical school in eastern Inis Mór at roughly the same time. This explains not only the presence of relatively Standard Irish variants in her speech (in contrast to the speech of her husband), but also specific traits which seem to link her to her age-mate from western Inis Mór. Thus the broader teen-age socialization patterns of younger generations may again explain these anomalies.

10. Some innovations of the youngest informant

Let us review briefly the innovations of the youngest informant on Inis Oírr. In the area of phonology, we have seen that the older male informant on Inis Oírr produces no examples of the affricate [t͡ʃ] in words such as teach ([t’æːχ]) or tinn ([t’iːN’]). His wife, the older female informant mentioned above, produces a modest number of examples of affrication, but the youngest informant produces twice that number of examples. Apparently the younger generation is moving strongly in the direction of affrication of palatalized dental consonants.

Turning now to morphology, in the case of the emphatic pronominal form seisean ([ʃεʃəN]), we find an isolated variant in eastern Inis Mór in which the masculine emphatic particle -san has been suffixed to the unemphatic masculine pronoun sé, producing sé-san ([ʃeːsəN]).

This same male informant in eastern Inis Mór uses the same kind of a construction in producing the third person singular feminine emphatic pronoun sí-se ([ʃiːʃε]), in which the feminine emphatic particle -se has been suffixed to the unemphatic feminine pronoun sí. He also produces the Standard Irish emphatic variant sise ([ʃɪʃε]), with a lax vowel in the first syllable. The male informant on Inis Meáin does exactly the same thing, producing both variants.

In contrast, the youngest female informant on Inis Oírr uses the same kind of construction to form an emphatic feminine pronominal form, but she suffixes the masculine emphatic suffix -sean ([-ʃəN]) to the feminine unemphatic pronoun sí, producing sí-sean ([ʃiːʃəN]). Apparently she has extended the masculine emphatic suffix to cover referents of both sexes.

In regard to the inflected prepositional pronominal form tharstu (‘past them’), the young female informant produces two innovative forms: the anomalous form thair ([hær’])Footnote ⁶ and the analytic construction thar iad ([har iːəd]). This latter construction is a typological innovation, since Standard Irish does not permit analytic constructions of preposition-plus-pronoun.

In the area of the irregular verbs, recall her positive and negative future forms of the verb tabhair, (‘give’): túire ([tuːr’ɪ]) and ní thúire ([n’iː huːr’ε]), respectively, where she palatalizes the flap in both forms.

Recall also her innovative variants for the negative preterite forms of the verb déan, (‘do’): níor dhín ([n’iːr jiːn’]) and níor dhion ([n’iːr jəN]). While the long high front vowel [iː] of the first variant can be heard in other forms of the paradigm for the same verb, the neutral vowel [ə] of the second variant is strikingly unique.

11. Innovative syntactic constructions

On Inis Oírr we also find two strikingly innovative syntactic constructions, both coming from informants who are a generation older than the young woman just discussed. The first construction, an interrogative copulative construction with an indefinite predicate, would be expressed in Standard Irish as An iascaire tú? (‘Are you a fisherman?’).

In this case, however, both husband and wife produce constructions in which the interrogative particle an is replaced by an invariable interrogative particle ab é containing the third person masculine singular pronoun é, a pronoun which is totally inappropriate in this context: ‘B é iascair’ ‘ú? and ‘B é iascaire tú/tusa?, respectively.

The second construction, an interrogative copulative construction with a definite predicate, would be expressed in Standard Irish as An tú Séamas? (‘Are you Séamas?’). In the responses of these two informants, however, again we encounter the same invariable interrogative marker ab, but now a variant of the second person singular pronoun tú is suffixed to it. Thus we get Ab iú Séamas? (with palatalization of the interrogative particle) from both informants.

Both these constructions violate basic rules of Irish syntax and are not encountered in the data from elsewhere in the islands nor from any mainland Connemara dialect.Footnote ⁷

The young woman next door, however, the youngest informant of our sample, shows no sign of this unusual syntax. Instead, she produces the Standard Irish constructions An iascaire tusa? and A’ tusa Séamas? Once again, though she herself innovates in other areas of the grammar, this particular rejection of the older copulative form illustrates the opposing tendency found in the youngest, most well-educated generation—the tendency to veer in the direction of Standard Irish.

12. Gender-related linguistic variation

Social activities in the Aran Islands outside of the family tend to be strongly segregated by sex, and this segregation of social networks by gender has left its mark on the phonology, the morphology, and the syntax of the local Irish language.

Examining again the affrication of the initial consonant in such words as teach, tinn, etc., we note that the female informant in easternmost Inis Mór affricates nearly all her examples, the female informant on Inis Meáin affricates all her examples, but the older female informant on Inis Oírr affricates only half of her examples. Nevertheless, the youngest female informant there produces twice as many affricated examples as non-affricated examples, suggesting a generational change and approximation to the speech behavior of the two western islands.

Turning now to the male informants, the male informant in easternmost Inis Mór produces nearly twice as many affricated examples as non-affricated examples, the male informant in Inis Meáin affricates less than half his examples, and the male informant in Inis Oírr produces no affricated examples at all.

Judging from the sharp decline in frequencies among male speakers as we move eastwards, it seems as though the innovation may have originated among female speakers in easternmost Inis Mór, and then spread eastward to Inis Meáin and finally to Inis Oírr. The two female speakers in the two western islands are roughly of the same age, but the older female speaker in Inis Oírr is a generation younger and the youngest female speaker in Inis Oírr is a generation younger still.

The recency of the innovation is suggested not only by its slow movement through the female generations in Inis Oírr but also by the fact that it has gotten no foothold in Inis Mór among either sex to the west of the easternmost area of the island. That easternmost Inis Mór may have been the region which exported the innovation and Inis Meáin the receiving region is further suggested by the relatively dense population in the townlands surrounding the harbor in easternmost Inis Mór vs. the sparse population in Inis Meáin.

A case of an innovation moving in the opposite direction might be that of a tendency to raise the vowel [æ] to [ε] before non-sonorants in words such as deas, teach, etc. Both the male and female informants in Inis Meáin exhibit that tendency, but it is only the female informants in eastern Inis Mór who show the trait. Remembering that marriage of women from the eastern islands and from mainland Connemara into Inis Mór has been frequent over the generations, it may be that the trait was introduced into eastern Inis Mór by women from Inis Meáin who passed the trait on to their female descendants. Still, it is also possible that a trait characterizing women in eastern Inis Mór was transmitted to women on Inis Meáin and generalized to both genders there on the smaller island.

One last phonological trait is worthy of notice. In Standard Irish, the voiceless glottal fricative or approximant [h] is inserted in phrase-level constructions by a sandhi rule to prevent adjacent vowels from coalescing at word boundaries. On Inis Mór, all the male informants elide or block the intrusive [h], producing examples such as Ná imígí! and Ná osclaígí! instead of Ná himígí and Ná hosclaígí. In contrast, all the women insert the [h], as in Standard Irish. Only the youngest female informant on Inis Mór has apparently borrowed the blocking rule from the men, a possible sign of generational change among the women on Inis Mór.

13. Gender-related morphological variation

The variant cé for the locative interrogative particle cá (Where…?) is widespread in mainland Connemara, but it is nearly absent on the Aran Islands. Only one older female informant in easternmost Inis Mór and her counterpart on Inis Meáin use the variant cé. Perhaps the variant was introduced into easternmost Inis Mór and Inis Meáin by women marrying into the Aran Islands from Connemara, but the variant has certainly not taken root on the islands.

Note also clause-final tú (‘thou’) in classificatory copulative constructions, such as Standard Irish An iascaire tú? (‘Are you a fisherman?’). All the informants across the islands lenite or elide the initial consonant of the pronoun, producing thú ([huː]) or ‘ú ([uː]), but four women, on Inis Mór and on Inis Oírr only, produce examples of unlenited/unelided tú, as in Standard Irish. In contrast, when the pronoun occurs as a direct object in a transitive clause, all informants on all three islands lenite the pronoun.

Finally, in regard to the prepositional pronoun tríthi, (‘through her’), we find a separate development of two sets of variants within separate male and female networks.

We find male informants on the two western islands producing variants based on the masculine pronominal form tríd (‘through him’) to refer to female referents, e.g., tríd, ‘roíd ([riːd’]), and the analytic construction tríd í. In contrast, female informants on all three islands produce an alternative variant tríti/tríte.

Thus the male informants have simply generalized the masculine pronominal form to female referents, while female informants have created a hypercorrect form which preserves clearly the distinctiveness of the feminine form.

14. Gender-related syntactic variation

There is in Standard Irish an interrogative construction of cause, which contains a head phrase Cén fáth…, followed by an indirect relative clause, e.g., Cén fáth a ndeachaigh [X]…? (‘Why did [X] go…?’). Some informants, however, substitute a subordinate clause headed by the particle go for the indirect relative clause, e.g., Cén fáth go ndeachaigh [X]…? Note that the insertion of the subordinating particle go is associated almost exclusively with female informants across all three islands. This is especially true of the youngest informant on Inis Mór, who inserts the particle regularly in interrogative constructions of cause and at least in one interrogative construction of manner: Cén chaoi a bhfuil…? > Cén chaoi go bhfuil…? (‘How is/are…?’). This may indicate a generational change among the women on Inis Mór.

A second example of gender-related linguistic traits is found in disjunctive interrogative copulative constructions embedded in a larger framework, e.g., Níl a fhios aige an é seo é nó nach é (‘He doesn’t know whether this is it or not’). Focusing on the final phrase … nó nach é, we find that Inis Mór is distinguished from the other two islands by the peculiar construction …nó nach b’ é.

This construction is used by nearly half the informants across Inis Mór, and of these four informants, three of them are women. The question as to whether the linguistic trait can be considered more characteristic of women is still open, however, since a related construction, the tag question …nach b’ ea? (Isn’t it?), was reported used by a second man from Inis Mór.

15. Individual creativity

We have already seen the surprising diversity of ways in which individual speakers distinguish the present tense forms from the future tense forms of the verb abair, (‘say’), but there are more prosaic examples of individual creativity.

There are a number of instances in the data where informants seem to be exploring a wide range of possibilities as they produce variants, and in certain cases, they seem to be exploring all logical possibilities. In such cases, it may be that none of the variants has yet been adopted by a social group as a recognized marker of that group.

We can start with an example of typological change in the morphology and syntax, where an object noun has been marked for the genitive case traditionally by inflection and lenition. In terms of geolinguistics, we find the Standard Irish construction …ag coinneáil Sheáin… (‘…keeping Seán…’) in western Inis Mór and in Inis Meáin, reflecting the greater conservatism of the two areas. Towards the center of Inis Mór, however, and all the way eastward, with a single exception, only non-standard variants are produced. We find the constructions …ag coinneáil Seáin…, with inflection of the noun but no lenition, …ag coinneáil Sheán…, with lenition of the noun but no inflection, and …ag coinneáil Seán…, with neither inflection nor lenition.

On Inis Oírr, both female informants produce the variant …ag coinneáil Seán…, while the male informant produces …ag coinneáil Seáin….

Thus, in terms of geolinguistics, both western Inis Mór and Inis Meáin preserve the traditional forms, while both eastern Inis Mór and Inis Oírr produce all the non-standard variants. While it seems impossible to assign a locality to the individual non-standard variants, it does seem possible to assign the non-standard variants by gender. The variant …ag coinneáil Seán…, with no inflection and no lenition, characterizes female informants strongly (four female informants as opposed to one male informant); the variant …ag coinneáil Seáin…, with inflection and no lenition, characterizes two widely separated male informants, and the variant …ag coinneáil Sheán…, with lenition and no inflection, characterizes a single female informant.

Head-words or head-phrases of dependent clauses are another area where a wide range of variants is often offered. For example, the head-phrase Cá fhad…? (‘How long…?’) can be used to form the temporal interrogative construction Cá fhad go rachaidh [X]…? (‘How long till [X] goes…?’). If we focus solely on head-words and head-phrases etymologically related to the word fad (‘distance, length of time’), we find the following list of variants: Cáide…? (a blended form), Cén fhad…?, Cén fhaid…? Cén t-ad…?, Cén faide…?, and Cé chomh fada…?. In this case, it seems as though the informants throughout the islands are still exploring the logical possibilities in regard to the form of the noun, including its gender assignment, and that no particular variant has yet been adopted by a social group as a group marker.

Another head-word which is characterized by a surprising range of variants, both on the Aran Islands and in the Cois Fharraige dialect on the Connemara mainland, is the head-word sula, (‘before’). Ignoring variants which are not etymologically related to sula, in the present tense we find the variants sula, chola, hlá, and hol, the compound forms sol má, sol ma and sol mar, and the parallel forms sór and sor.

In the simple past constructions, we find Standard Irish sular along with the variants solar, sola, ‘lar, sulár, suláir, holáir, and soláthar. We also find the parallel series sorar, soirear, sor mar, and sar.

It is difficult to assign these variants to different areas of Inis Mór or to separate islands. The answer may lie in the fact that these head-word or head-phrase variants are relatively low-frequency forms, compared to the irregular verbs and pronouns we have seen already, which can be characterized as high-frequency forms in everyday speech. In other words, it may be that the head-words or head-phrases of temporal clauses are simply not frequent enough in the speech of the playgroup to be selected as group markers.

A final example of this exploration of all logical possibilities among informants is the variation in the dependent clause Dá mba bhean thú… (‘If you were a woman…’). In addition to the Standard Irish construction shown above, with eclipsis of the copula and lenition of the predicate noun, we find:

Though it is difficult to assign these syntactic variants to specific localities, age groups or genders, once again virtually all logical possibilities have been exploited in roughly equal measure by informants across the islands in order to produce these variants.