How much L1 does a terminal speaker know?

Edge Hill University BA (Hons) English LNG3129 Language Dissertation

HOW MUCH L1 DOES A TERMINAL SPEAKER KNOW?

By: Gemma Simpkin

Supervisor: Professor Anthony Grant

May 2014

1 Abstract There are theories that as a language becomes obsolescent, some features of that language begin to erode as the number of speakers decreases. A linguist conducting fieldwork in a moribund language needs to consider how to elicit the most language from their consultant in the time that they have. Terminal speakers of a language will be able to communicate in that language to different levels of fluency. This dissertation looks at three different cases of language death, and in particular, the fluency and knowledge of the last speaker of each of these languages. There are various methods of testing how much a terminal speaker knows, such as looking at their ability to use complex sentences, looking at their vocabulary, and in particular, as Dorian (1977:29-30) suggests, looking at previous studies in order to evaluate how the language has changed. It was found that although different types of terminal speaker can be categorised, it is not possible to generalise how much terminal speakers know of their languages; it varies based on the individual. However, one of the most significant factors affecting how much a terminal speaker knows is the period of time between where they ceased using their language to communicate with other speakers, and the time that they became a consultant of their language. Keywords: Manx, Wappo, Tunica, language death, terminal speaker, fluency, elicitation, sentence complexity, vocabulary, disuse. Acknowledgements Most importantly of all, I would like to thank Professor Anthony Grant for supervising this dissertation and being a great help in obtaining the information I required to complete it, and also for providing me with prompt and excellent advice throughout the writing process.

2 Secondly, I would like to thank my partner Ryan Garner for his help in allowing me to access some of the literature I needed to complete this dissertation. I would also like to thank him for his efforts in proofreading my work. Thirdly, I would like to thank my father, Philip Simpkin in his efforts to aid me with obtaining information for my dissertation, which I would not have found without his help. In addition to this, I would like to give a special thanks to Brian Stowell for providing me with valuable information about Manx.

3

Table of contents Page List of tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

List of figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

List of images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

6

Chapter 1: The case of Manx and its last fluent native speaker, Ned Maddrell 1.1 The history and decline of Manx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9

1.2 English influence on Manx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

1.3 Edward ‘Ned’ Maddrell’s (1877-1974) fluency in Manx . . . . . . . . . . . . .

13

1.4 Sentence complexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

1.5 Features of Late Spoken Manx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

19

1.6 English influence on Ned Maddrell’s Manx . . . . . . . . . . . . . . . . . . . . . . .

19

1.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

Chapter 2: The case of Wappo and its last fluent native speaker, Laura Fish Somersal 2.1 The history of Wappo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

2.2 Dóloris ‘Laura’ Fish Somersal’s (1898-1990) fluency in Wappo . . . . . . .

25

2.3 A study of Wappo animal terms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

27

2.4 Sentence Complexity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

2.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

31

4 Chapter 3: The case of Tunica and its last fluent native speaker, Sesostrie Youchigant 3.1 The history and decline of Tunica . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32

3.2 Sesostrie Youchigant’s (c.1870-a.1940-50) fluency in Tunica . . . . . . . .

34

3.3 Knowledge of lexical items . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

36

3.4 Tunica Texts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

3.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

38

Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

39

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

Web links . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

44

5

List of tables Description

Page

1. The distribution of Manx and English speakers in Cregneash in 1901.

15

2. Sentence complexity in Ned Maddrell’s Manx.

18

3. Lengths of time in which each speaker’s L1 was not used.

40

List of figures Description

Page

1. Map of the Isle of Man, including parish divisions.

10

2. Manx text accompanied by an English translation.

17

3. Map of indigenous languages of California at the time of European contact.

22

4. Map of the Yuki-Wappo tribes.

24

5. Different species of animal in Radin and Sawyer’s vocabularies of Wappo.

28

6. Map of pre-contact Southeastern American tribal languages, including Tunica.

33

List of images Description

Page

1. Ned Maddrell

13

2. Laura Fish Somersal

25

3. Sesostrie Youchigant

34

6

Introduction To assess how much L1 a terminal speaker knows, there are numerous factors which must be taken into account when studying individual speakers. In this paper, three case studies of three terminal speakers are presented. Ned Maddrell, the last known speaker of the Manx language of the Isle of Man is the first case to be presented. The second is Laura Fish Somersal, the last known speaker of the Wappo language, once spoken in the Alexander Valley in California. The third case is Sesostrie Youchigant, the last known speaker of the Tunica language, spoken in Marksville, Louisiana. An important consideration when looking at the fluency of an individual is looking at their background. The person or people who they learned their language from, their upbringing and their level of education all contribute to their level of fluency. The period of disuse of their first language also dictates their level of fluency. Youchigant had the longest period of disuse of the three speakers, having not spoken Tunica in eighteen years since the death of his mother, which is reflected somewhat in his recollection of the language (Haas 1953:175). Other factors such as the weak areas in the speaker's lexicon, the loss of lexical or grammatical features when their material is compared to older studies or sources of the language, and their ability to innovate and construct complex sentences also indicates how proficient each speaker was with their first language.

A terminal speaker is defined by Tsitsipis (1989: 119) as a speaker who has heavy lexical losses, loss or confusion of phonological oppositions, simplification of grammar and who communicates through the use of formulaic expressions. The term 'semi-speaker' is also seen to be used interchangeably with 'terminal speaker' in literature as noted by Quesada (2000). Campbell and Muntzel (1989:181) classify speakers of dying languages into four

7 categories. S or “strong” refers to almost fully competent fluent speakers, I or “imperfect” refers to reasonably fluent semi-speakers, W or “weak semi-speakers” refer to speakers which more restricted competence, and R or “rememberers” refers to those who only know a few isolated words or phrases of their language.

An example which can be classified using this scale are the terminal speakers in Tsitsipis' work involving Arvanítika, a variety of Albanian which is spoken in Greece. Tsitsipis (1989:120-27) notes that the terminal speakers of Arvanítika ranged from under 20 years old to 40 years old. Their use of the language was restricted to when they needed to clarify something for an older speaker, or to impress a foreigner. Younger terminal speakers were seen to use formulaic expressions which were more culturally appropriate for older, female members of the community to use. It can therefore be gauged that Tsitsipis' terminal speakers would fall towards the lower end of Campbell and Muntzel's “weak semispeakers” category, almost verging on rememberers.

The correlation between the chronology of language and speaker fluency is another significant factor to consider. Ned Madrell went to live with his aunt Paaie Humman, or Margaret Taubman if her name is anglicised, between the ages of two and a half to five. His aunt died when he was thirteen in 1890 (Broderick 1984:467-8). Laura Fish Somersal was known to have conversations with a sister in Wappo until the mid-1960s, and showed “a relative ease in making up new words” according to Elmendorf (1981:39). Elmendorf also states that due to Somersal's lack of schooling and restricted language use with only her family, she only has a single level of language use, but demonstrates a high degree of fluency of that level. Sesostrie Youchigant was exposed to the language by his mother who spoke to him in Tunica. However, he usually replied in Louisiana French, according to Haas (1940:9). This may therefore have restricted his ability to innovate using Tunica.

8

Language death is seen as a pervasive problem for the worlds' languages. Ethnologue lists 7105 world languages as of 2013, of which approximately 33.6% are categorised as endangered or dying. However, the figure which Ethnologue gives cannot be wholly accurate, as many linguists disagree on what constitutes a language and what constitutes a dialect. According to McGregor (2009: 301-2) languages can be defined using three different methods, or 'senses'. The political sense of language definition is defined as “the forms of speech that are associated with nations”. The second type of language definition which is commonly used by linguists is the mutual intelligibility criteria. The third type of language definition is the idea that if in forms of speech, there is a certain percentage of common or similar words and morphemes, it can be defined as a language. For example, Hindi and Urdu are classified as separate languages. Both are largely mutually intelligible in spoken form, but what separates them is that they are divided politically, culturally, and orthographically (Fennig 2013).

There are numerous reasons why languages die. Campbell and Muntzel (1989: 182-6) establish four kinds of language death situations; sudden death, radical death, gradual death and bottom-to-top death. Gradual language death is is the type of language death which most of the related literature deals with. It is a loss of language due to a shift in the dominant language of the area. Manx, Wappo and Tunica are all examples of gradual language death. English became the dominant language over all three of the above languages, although Louisiana French was the dominant language in the area over Tunica at one point. This later gave way to English, as Haas (1940:10) mentions that Youchigant is polylingual in Tunica, Louisiana French and English. However, in the case of Tunica, other indigenous languages in its area also affected its decline.

9

1. The case of Manx and its last fluent native speaker, Ned Maddrell 1.1 The history and decline of Manx The Manx language is a Celtic language of the Goidelic branch, which was spoken as a first language on the Isle of Man from around 300AD until the death of its last native speaker, Ned Maddrell, in 1974 (Broderick 1993:228). In the fourth and fifth centuries, Irish settlers came to the Isle of Man, displacing the language which previously existed there, which was probably a Brythonic language according to Stowell (2005:384). Manx diverged from Irish around the thirteenth century and from Scottish Gaelic around the fifteenth century. Manx shows some features which are obsolete in the other Goidelic languages, but it also has some innovations which separate it from Irish and Scottish Gaelic. Its relative isolation from its sister dialects allowed it to develop on its own, leading Manx to innovate in areas such as its verbal system (Broderick 1993:229). Broderick defines three periods of Manx; Early Manx in the seventeenth century, Classical Manx in the eighteenth century, and Late Manx in the nineteenth and twentieth centuries. The latter period was the period in which the Manx language was fading out of use in favour of English. Figure 1 below demonstrates the distribution of Manx in 1875.

10

Figure 1: Map of the Isle of Man from Jenner (1875), including parish divisions. The shaded segments represent primarily Manx-speaking areas as of 1875. A key for the labelled parishes, towns and other notable areas are shown below.

1. Bride 2. Andreas 3. Lezayre 4. Michael 5. Ballaugh 6. Jurby 7. Lonan 8. Maughold 9. German 10. Patrick 11. Marown 12. Rushen

13. Arbory 14. Malew 15. Santon 16. Braddan 17. Onchan I. Dalby II. Cregneash C. Castletown D. Douglas P. Peel R. Ramsay S. Spanish Head

11 1.2 English influence on Manx As the dominant language of the Isle of Man shifted from Manx to English, Manx was affected by these changes. Broderick (1984:7) highlights some loss of features from Classical Manx (CM) to Late Manx in what he calls ‘Late Spoken Manx’ (LSM or LM), spoken from the early 1800s-1974. Classical Manx is the period in which Broderick believes the Manx language began to change due to the equilibrium between Manx and English on the Isle of Man being upset (Broderick 1999:80). Broderick notes that some grammatical and phonological features have broken down somewhat since the period of CM in the eighteenth century. Manx, like Irish, had opposition between broad and slender consonants, but by the LM period, this system had broken down and became erratic. Kewley Draskau (2008:223) defines a slender consonant as a palatalised consonant, and in Present Day Manx, palatalisation occurs in final position, or when /t, d, l/ occur after /i, e/. Grammatically, consonant lenition in LM was also “in a state of flux”, with the rules regarding when lenition occurs becoming irregular. Another feature of LM is the disappearance of the oblique cases and the gendered and synthetic form of verbs. The loss of gender may be a resulting factor of English influence, as English had lost its gendered inflections in the Old English period (Baugh & Cable 1993:154). As more parents neglected to teach their children Manx, the speakers later may have had difficulty acquiring features of Manx which were not present in English. Kewley Draskau (2005:229) cites O’Rahilly, who argued that English “played havoc” with Manx syntax since the beginning of its life as a written language, and that some printed Manx is merely English disguised in Manx vocabulary. Moore (1895-1901) looked at the amount of English-origin words in Manx by studying Reverend John Kelly’s Manx dictionary published in 1866, stating that the letter ‘p’ had thirty words of native origin, and the rest, about 300 words, were of English origin. Kewley Draskau highlights another feature of LM; the speakers would use periphrastic verb tenses rather than inflected ones. It is possible to avoid the

12 use of inflected tenses completely in Manx, using ve ‘be’ and jannoo ‘do’ as auxiliaries. Broderick (1984:82) noted that jannoo mainly became used as an auxiliary in LM. Kewley Draskau adds that in CM, regular preterites in narrative discourse was a standard occurrence, but in LM, confident production of the inflected forms decreased, leading to speakers of LM limiting their use of inflected forms to only the irregular verbs.

13

1.3 Edward ‘Ned’ Maddrell’s (1877-1974) fluency in Manx

Image from Greinneyder, Flickr

14 Kenneth Jackson visited the Isle of Man in 1950-1 and described the 72 year old Ned Maddrell (NM) as the youngest, most alert and most fluent surviving speaker of Manx (Jackson 1955:vi). Broderick (1999:74-5) also believes that NM was much more fluent than early recordings of him suggested. NM was born in 1877 and learned to speak English before he spoke Manx, as his parents refused to speak Manx in front of their children (Broderick 1984:463-8). He learned Manx from his great aunt, Paaie Humman (or Margaret Taubman1) at the age of five when he went to live with her in Cregneash, Rushen. However, Broderick (1999:75) states that NM went to live with his aunt at two and a half years old because of the size of his family. As NM learned Manx from his aunt, it is important to know which generation of Manx speakers Paaie Humman was born into. The Isle of Man Family History Society (IoMFHS) states in their burial records that a Margaret Taubman of Rushen died in 1890 at the age of 82, which corresponds to NM’s account that she died when he was 13 years old (Broderick 1984:468). This would make her birth year around 1808. According to evidence from NM’s texts, his aunt seemed to be either a Manx monoglot or had a very limited knowledge of English. This would not have been unusual at the time in Cregneash, where many sources state that Manx in Cregneash may have lingered into the first decade of the twentieth century (Miller 2007, Broderick 1999:5, Jackson 1955:1). The dates given for the shift from Manx to English in the Isle of Man as a whole are 1840-80, with the last generation of children receiving Manx from the crib in the 1860s-80s (Broderick 1993:283-4). Broderick (1999:24) defines the late eighteenth century and early nineteenth century as a period of mass migration both to and from the Isle of Man, which contributed to the equilibrium between Manx and English bilingualism being upset in favour of English. Paaie Humman would have been of the last few generations to learn Manx natively and monolingually. Her The surname ‘Taubman’ is speculated to be of Dutch origin. Descent can be traced back to King Edward I of England. Taubmans have been found in high positions in the Isle of Man since the 18 th century, with at least three members of the family having been elected as speakers of the House of Keys, which is the lower house of the Manx parliament Tynwald. (Moore 1889). 1

15 Manx may not have been influenced greatly by English due to the minimal contact of English and Manx in Cregneash until the second half of the nineteenth century. It can therefore be speculated that NM may not have had quite as many English-influenced features in his Manx as other speakers may have had, since he learned a variety of Manx that may have been moderately isolated from English. Jackson’s oldest informant, John Kneen, ‘the Gaaue’ (the Smith) of Ballaugh was born in 1852, and was aged 97 in 1950 (Ballaugh Heritage Trust 2008, Jackson 1955:3). Kneen would have been two generations younger than Paaie Humman, and may not have had access to the community Manx in Ballaugh that NM would have had access to in Cregneash. The significance of Cregneash itself should not be underestimated when considering the variety of Manx that NM would have learned. In 1901, a census was carried out, which showed that although Manx had ceased to be a language taught to children in other areas by 1880, the language had been taught from the crib for longer in Cregneash than in other places. Table 1: The distribution of Manx and English speakers in Cregneash in 1901 Age group

Number of people

Manx (44)

English (45)

1 to 20

39

3

36

20 to 40

17

8

9

40 to 60

23

23

-

60+

10

10

-

(Oldest = 75)

(Oldest = 38)

From Miller, S. (2007) Here the Manx language lingers, and may linger some time longer: Manx and English in Cregneash in 1901.

Table 1 shows that although there is an almost equal distribution of Manx and English

16 speakers, the difference in language use between the age groups showed that English was pushing Manx out in the younger generation by this time, whereas the older speakers showed a very clear preference towards Manx. An important thing to consider is how NM’s life at sea and his age when he became a consultant may have affected his Manx. NM’s aunt died when he was thirteen years old. He had gone to school in Port St. Mary, and left at this age to go to sea. Many of NM’s stories collected by Broderick (1984:462-72) involve his life at sea. He may have managed to keep hold of some of his Manx by conversing with Gaelic speaking sailors, as there is some degree of mutual intelligibility between the languages (Miller 1993). One notable story highlighting this mutual intelligibility is when NM’s crew were buying provisions in Sutherland, Scotland, a woman selling eggs mistook the crew for Englishmen and attempted to overcharge them, but the Scottish Gaelic speakers could be understood by the Manx skipper (Broderick 1984:464-5). NM stayed at sea most of his life, retiring at age 70 (Mac Reachtain 1993:129). The term ‘rusty speaker’ coined by Lise Menn has been used by Broderick (1999:6) to define many of the last native speakers of Manx in their later years. NM himself has discussed that his Manx was not as good as it once had been, stating that he could once speak Manx better than he could English (Broderick 1984:468). In order to assess the fluency of NM, one of the texts which he provided shown in Figure 2 will be analysed, with particular focus on sentence complexity, the presence of typical features of Late Spoken Manx, and English influence on NM’s Manx. The text in Figure 2 also demonstrates the role of Manx as a community language in Cregneash, as opposed to English in the majority of the Isle of Man.

Figure 2: Falling down the stairs by Ned Maddrell taken from Broderick (1999:274).

17 Manx 1

Va mee ruggit ayns y Corvalley, as tra va mee mysh blein dye ash va mee curt lesh dys Creneash,

as un laa tra va mee mysh tree vleeantyn dye ash va mee curt lesh marish my vumming son dy cur shilley er chenn naunt jeh my vumming, as va mee …, er-lhiam dy row mee feer mitchooragh, as va mee goll mygeayrt as cha’s aym c’red va mee jannoo ooilley, as snaue mee 5

seose ny greeishyn. Tra va mee ec y vullagh ny greeishyn va mee laccal fosley yn dorrysh jeh nane

jeh ny shamyryn, agh ren mee coayl my holt as tuitt mee sheese ny greeishyn. Va’n chenn ven soie ec y bun ny greeishyn er stoyll as ren ee clashtyn yn sheean as roie ee gys ny greeishyn as cum ee magh e vrat as ren me tuittym ayn. As ren ee cur lesh me gys my vummig as dooyrt ee, - Shoh gow eh! Lurg shen va me curt lesh lesh my stroin as y fuill roie ass gys ven elley va jannoo 10

oalys son dy freayll yn uill, agh cha’s aym, cha nel cooinaghtyn aym c’red ren taghyrt lurg shen, agh

veagh sheih briaght jeem, - C’red ren Auntie Paaie gra? As yinnin gra, - Shoh gow eh! As ta shen yn chied cooinaghtyn t’aym’s jeh Gaelg, as ren eh taghyrt ayns Creneash tra va mee mysh tree vleeantyn dy eash!

English I was born at the Corvalley, and when I was about a year old I was brought to Cregneash and one day when I was about three years old, I was taken by my mother for to visit an old aunt of my mother’s. And I was … I think I was very mischievous, and I was going around and I don’t know all of what I was doing, but I crawled up the stairs. When I was at the top of the stairs I was wanting to open the door of one of the rooms, but I lost my grip and fell down the stairs. The old woman was sitting at the bottom of the stairs on a stool and she heard the noise, and she ran to the stairs and held out her apron and I fell into it. And she brought me to my mother and said, - Here, take him! After that I was taken with my nose bleeding to another woman who was making a charm for to stop the blood, but I don’t know, I don’t remember what happened after that, but people would ask me, - What did Auntie Paaie say? And I would say, Shoh gow eh! And that’s the first memory I have of Manx, and that happened in Cregneash when I was about three years old.

1.4 Sentence Complexity Sentence complexity is a good indicator of the fluency of a terminal speaker. According to Voegelin & Voegelin, (1977:333-8) endangered languages go through a process of ‘deacquisition’, where simplification of the minority language occurs due to influence exerted by the dominant language. In the Uto-Aztecan language Tübatulabal, a generational shift in which the newer generation of native speakers had more difficulty producing or

18 interpreting complex sentences took place. Voegelin & Voegelin relate the younger generation’s lack of proficiency in producing complex sentences to child language acquisition theories that the ability to process complex sentences is learned by children by their teenage years. Carl Marstrander visited the Isle of Man in 1929, 1930 and 1933 to look for speakers of Manx. One criteria which he used to assess speaker fluency was the production of complex sentences. He notes two men from Cregneash, the youngest being 69, were able to produce simple and compound forms, but could not translate a complex sentence which was given to them (Broderick 1999:208). Hill (1978:56-7) expands on this struggle to produce complex forms through studying a Uto-Aztecan language of California; Cupeño. In Cupeño, speakers were observed using complex sentences much more frequently in 1920 than in 1962. Hill argues that the frequency of production of complex sentences at different points in the languages’ life can be an indicator of language decay. The men interviewed by Marstrander would have been born at least 17 years earlier than NM and do exhibit this marker of language decay as described by Hill. In Figure 1, NM produced eight sentences. Table 2 demonstrates the distribution of sentence types in NM’s speech. Table 2: Sentence complexity in NM’s Manx Sentence Type

Number of occurrences

Percentage

Simple

1

12.5%

Compound

2

25%

Complex

0

0%

Compound-Complex

5

62.5%

Total:

8

100%

19 Although there are no solely complex sentences in NM’s speech, the majority are compound-complex. Subordination still occurs in NM’s speech, so given the opportunity he would be able to competently produce complex sentences. Compound sentences also occur a quarter of the time, and the coordinating conjunction as ‘and’ is used frequently throughout the text. Looking at NM’s Manx, it does not suggest that he has acquired features of language de-acquisition associated with terminal speakers of a language. NM would have still been speaking Manx as his primary mode of communication at the age that children would learn to process complex sentences. The two men from Cregneash who were interviewed by Marstrander, although being considerably older than NM, were unable to produce a complex sentence upon elicitation. It may be the case that the manner of elicitation between Marstrander’s informants and NM’s text above are different. However, it seems to be that after the passing of Manx as a community language in Cregneash by 1910, (Broderick 1999:44) many other less fluent speakers of Manx who were raised bilingual or in some cases, with just English in their homes would not have been exposed to these complex forms as regularly as more fluent speakers such as NM might have. 1.5 Features of Late Spoken Manx NM is seen to use periphrasis of the verb in some parts of his speech. An example of this is as ren eh taghyrt ayns Creneash (12) lit. ‘and it did happen in Cregneash’. In Figure 2, NM uses ren, the preterite of jannoo, along with the verb-noun taghyrt, ‘happen’. This example in its inflected form would be as haghyr eh ayns Creneash (Kewley Draskau 2008:118, 122-3). 1.6 English influence on Ned Maddrell’s Manx Broderick (1999:158) states that English exerted such a large influence on Manx that native Manx speakers would often calque English in their speech, leading to uncertainty and loss of idiom over time. One example which Broderick gives is that in LM, English but

20 would be used as a coordinating conjunction instead of the native Manx agh. NM does not do this, as he is shown to have used agh (6, 10, 11) as one of his coordinating conjunctions. Another feature of NM’s Manx which is not influenced by English is in his use of the definite article. In CM, the definite article would be written as followed: Singular (masc): y, yn Singular (fem): ny Plural: ny By LM, the feminine form had fallen into disuse aside from in fossilised phrases, due to the simplification of the Manx gender system. According to Broderick, (1999:120-1) speakers of LM, especially ones which had not used the language in a long time, would use the singular masculine definite article y or yn to refer to both singular or plural nouns. NM shows the contrary; although he does not use the feminine definite article, he does distinguish singular and plural forms, such as y vullagh (5) ‘the top’ and ny greeishyn (5) ‘the stairs’. NM does use English-influenced genitive forms, such as ec y bun ny greeishyn (7) ‘at the bottom of the stairs’, which would have been ec bun ny greeishyn lit. ‘at bottom of the stairs’ in CM. When NM’s aunt is referred to in the text, it is as both naunt and Auntie (11) which are both loanwords from English. Naunt has a gaelicised spelling, but is recognisable as a loanword from English. Kelly’s dictionary, (1866:209) finished in 1808 but published in 1866, cites shuyr ayrey ny moir as another term for ‘aunt’. This term does not appear in Cregeen’s 1835 Manx dictionary. Cregeen (1835:120) cites an attestation of naunt from the Manx translation of the Bible in Leviticus (18:14). The whole Bible had been translated into Manx by 1772, so naunt would have likely been a word originating in at least CM (Wood 1896). The borrowing of naunt may have been for simplification of kinship terms such as the term in Kelly’s dictionary. According to Tadmor, (2009:64) the semantic field of kinship has a

21 minor degree of susceptibility to borrowing from other languages, with a 15% borrowing rate. The use of naunt and Auntie reflects the influence that English would have had on not just LM, but CM also. However, it is notable that NM demonstrates knowledge of the older CM form of ‘aunt’ as well as other more recently borrowed forms. 1.7 Conclusion It is apparent that NM retained a high degree of fluency throughout his life, even though he had considered himself to be a rusty speaker. NM’s Manx is notable in that it was a relic of the variety of Manx that would have been used in areas like Cregneash two generations before his birth. Paaie Humman had a significant role in shaping NM’s Manx, and the fact that she was likely a Manx monoglot meant that NM’s Manx would have contained less Anglicisations than other Manx speakers. NM’s Manx contained more features of LM that had already emerged in the early nineteenth century in largely monolingual Manx areas. NM demonstrated proficiency in producing sentences containing subordinate clauses, and retained some native Manx features which had dropped out of use in other speakers of Manx, such as the use of agh rather than but. The CM distinction of the definite article plural was also a feature which other speakers of LM did not exhibit. An important consideration is that in some recordings of NM such as Jackson’s mainly elicited lexical items. Jackson noted that many of his informants struggled to recall various terms, and were apparently “unable to give the Manx for the commonest things” (Jackson 1955:3). Although this was attributed to the age and non-fluency of the speakers, it may be the case that some speakers demonstrate a better command of the language in spontaneous speech. NM certainly would have been a case of this, as the texts we have of him are the best indicator of his fluency. Above all, NM provided an excellent insight as to what Manx spoken as a community language would have been like in the early nineteenth century.

22 2. The case of Wappo and its last fluent native speaker, Laura Fish Somersal

Figure 3: A map of indigenous languages of California at the time of European contact, with Wappo underlined.

Image taken from Mason, B.J. (n.d.) California Indians Peoples

2.1 The history of Wappo The Wappo language was spoken in a twenty by fifty mile area of the Alexander Valley in

23 California until the death of its last native speaker, Laura Fish Somersal (LFS), in 1990 (Sawyer 1991:15). Wappo is classified as a Yukian language, also referred to as YukiWappo, although its legitimacy as a Yukian language is questioned by Sawyer (1977) due to divergence of Wappo from the other Yukian languages. Elmendorf (1968:181) shows that on the 200 word Swadesh list, Wappo has an average of just 26.5% cognates with two of the Yukian dialects, Huchnom and Yuki. However, Sawyer’s view that Wappo is unrelated to Yukian is not widely accepted. Yukian has been accepted as a language family since the 1870s. According to Golla, (2011:188) links have been suggested between Yukian and the Penutian, Siouan and Gulf families, but as of yet, it has not been affiliated with any larger language family. The name ‘Wappo’ apparently comes from Spanish guapo, ‘handsome, brave’, due to 19th century conflicts between the Spanish and the native population (Thompson et.al 2006:xi). Barrett described Wappo as having four mutually intelligible dialects in 1908, but by the 1950s when Jesse Sawyer began his work on Wappo, Wappo was described as having only two dialects (Elmendorf 1968:177). On Figure 4, Wappo is subdivided into five different zones, representing dialectal or cultural Wappo groups. These groups are the Mishewal, which LFS belonged to, the Guenoc or Locoma, a subgroup of Wappo which was the bilingual borderland between the Wappo and the Lake Miwok tribes, and the Mayacama, the Caymus and the Lile’ek, the latter of which was a group of Wappo who migrated northwards around 1800. Only the Mishewal and the Mayacama have reliable ethnographic and linguistic data (Golla 2011:191).

24 Figure 4: The Yuki-Wappo tribes

Map taken from Golla (2011) California Indian Languages p.189

25

26 2.2. Dóloris ‘Laura’ Fish Somersal’s (1898-1990) fluency in Wappo

Image taken from Dry Creek Rancheria (2014)

27

Dóloris ‘Laura’ Fish Somersal (LFS) was born in 1892 to a Wappo mother, Mary Ely, and a Southern Pomoan father, Bill Fish, near Geyserville, California (Sawyer 1976:121, AbelVidor et.al 1996:63).She would have spoken the Mishewal dialect of Wappo (Elmendorf 1981:277). LFS grew up bilingual in both Wappo and Southern Pomo as a child, but used Wappo much more often (Thompson et.al 2006:xiii). LFS did not attend school as she needed to care for her blind mother. She therefore could not read or write, and knew only how to sign her own name. She claims to have not learned English until she reached her late teens. She apparently held conversations in Wappo on a regular basis until the death of her sister in the mid-1960s (Elmendorf 1981:277). The linguist Jesse Sawyer worked with LFS from around the late 1950s until his death in 1986 (Shepherd 1991:5). She therefore would have still been using Wappo with her sister at the time that Sawyer began to work with her, so unlike the other cases presented in this paper, she did not have a long period of disuse or reduced contact with her L1. In later life, LFS became a basket weaver, and taught both basket weaving and Wappo at various universities and educational centres (Abel-Vidor et.al 1996:63-65). She died aged 92 in 1990. According to Elmendorf, (1981:277) LFS was not classified as an extreme case of a terminal speaker. Sawyer (1991:12) states that LFS’s English often contained calques from Wappo, which contrasts with the profile of a typical terminal speaker, whose L2 tends to exert a larger influence on their speech in adult life. Elmendorf (1981:277-80) stated that LFS’s lack of schooling restricted her use of language to one level, although she showed a high degree of proficiency at that level. On the other hand, LFS’s lack of schooling would have been a large factor contributing to the retention of her Wappo throughout her life. There is some evidence that a degree of morphological, lexical loss and stylistic loss has occurred from LFS’s mother’s generation to her own. Paul Radin studied Wappo in 1917 using two male informants, Jim Tripo and Joe McCloud. Both of Radin’s informants were

28 one generation older LFS, and when presented with the materials that Radin collected from them, LFS often could not understand the texts (Thompson et.al 2006:xii). This could either be because Radin’s phonology was unrepresentative of Wappo, or because LFS’s Wappo had changed. Tripo knew a form of ‘high language’ used for storytelling or myths, which LFS had no knowledge of (Elmendorf 1981:276). LFS disliked giving texts according to Thompson et.al (2006:xiii) as she felt she was “not qualified” to tell them, which may allude to the storytelling register known by Tripo. Driver (1936:179) also used Tripo as an informant regarding Wappo ethnography. Tripo lived with LFS’s mother, being described as a ‘probable cousin’ of her, which would make him a relative of LFS. As the advanced age of many terminal speakers of languages raises the issue of whether the memory of the informant is wholly reliable, Sawyer (1991:13-15) notes that LFS possessed a good memory. Although she had some lapses and occasionally mixed up Wappo and Southern Pomo words, she had a good ability to recall mistakes which she had made from months before without being prompted, and demonstrated an ability to make up new words with relative ease. 2.3. A study of Wappo animal terms Some animals hold some cultural importance to the Wappo. According to Driver (1936:186-199), the Wappo observed many taboos, often associated with women. Snakes were believed to cause rheumatism and even death in menstruating women, and a particular species of hawk, nihlekh, was believed to cause newborn babies to sicken or die. Many insects such as grasshoppers, caterpillars, snails and lice were once eaten by the Wappo, which, although perhaps not eaten by the time LFS was born, she may have still inherited the knowledge of the terms for these animals. Another notable animal mentioned in Driver (1936:217) is a Wappo deity, ‘old man coyote’, who was prayed to often while fishing or hunting to bring them luck. According to Elmendorf (1981:277) LFS acknowledged that she had some weak areas in

29 her lexicon such as zoological terms, which Elmendorf attributes to a gender divide. In order to investigate the alleged gaps in LFS’s knowledge, a comparison between the data which she gave and the data from Radin’s informants have been compared. The vocabulary provided by LFS is documented in Jesse Sawyer’s 1965 English-Wappo vocabulary, and it consists of around 2000 lemmas. In Sawyer’s unpublished Wappo field notes (1959-60), there is a paper named words not known or not remembered by Laura Somersal which consists of 27 words, 17 of which are related to animals in some way. In 1929, Radin published A Grammar of the Wappo Language, which included a dictionary, part of which consisted of approximately 527 nouns. Using this dictionary of nouns, Chart 1 below shows the occurrence of animal terms in both Radin and Sawyer’s vocabularies of Wappo, divided by animal species. The category ‘sea life’ has been expanded from just being ‘fish’ to encompass crustaceans and aquatic mammals such as sea lions.

30 Overall, 96 zoological items were found in Radin’s vocabulary, although one, transcribed as ts.i’ɔnɔl by Radin, is omitted from the table due to uncertainty about whether the word refers to a species of bird or a bumble bee. In Sawyer’s vocabulary, 108 zoological items were found. Figure 5 shows that in Radin’s vocabulary, birds made up the majority, whereas the vocabulary that LFS produced trends slightly higher in insect, mammal, sea life and amphibian terms. However, LFS’s vocabulary of birds is more advanced than what appears to be her strongest area of zoological lexicon, mammals. The reason for this is that an estimated 37% of all mammal terms in Sawyer’s vocabulary are loans from Spanish. Most of these loans relate to agricultural animals such as cows, sheep, donkeys, as well as mammals such as monkeys and hippos. In Radin’s vocabulary however, there is only one loan from Spanish included, ‘sheep’. This is likely due to the fact that Radin focused on recording native Wappo words while omitting most loans from Spanish. As a result, distribution of mammal terms may be skewed slightly. There are more words in Sawyer’s work that LFS is uncertain about at 3.7%, than words in Radin’s work that Tripo and McCloud may have been uncertain about at 1%. Comparing both works, it is shown that 32% of zoological terms which appear in Sawyer, excluding loanwords, are not seen in Radin, while 39% of zoological terms seen in Radin do not appear in Sawyer. There are some discrepancies which can be found between Sawyer and Radin’s works however. In order to address these discrepancies, the naturalist C. Hart Merriam’s c.1898 Miyakma dictionary has been consulted for his knowledge of zoological and botanical terms (archive.org 2014). One discrepancy, ‘panther’ in Radin, is recorded as katama, while in Sawyer, it is wima. According to Merriam, katama corresponds to mountain lion, which is another term for a panther while wima corresponds to a ‘big wolf’. On the other hand however, a word in Sawyer’s work which did correspond to Merriam’s work, where Radin’s didn’t, is the word for ‘gopher’. In both Merriam and Sawyer, ‘gopher’

31 was recorded as čoya, whereas in Radin, it was nete. Čoya for Radin, was ‘rat’. In Merriam, nete was ‘mole’, so Radin’s translation of čoya was incorrect in this case. Altogether, there were only approximately three words in Merriam’s dictionary which were not found in either Sawyer or Radin’s; antelope, badger and salamander, which suggests that all informants in both Sawyer and Radin’s works managed to produce as much as they could remember. 2.4. Sentence complexity Elmendorf stated that LFS showed an ability to produce sentences with embedded subordinate clauses with relative ease (1981:277). In Thompson, Park & Li’s Reference Grammar of Wappo (2006:xiv), it is stated that the grammar is not one of a speech community, but rather of one individual, LFS; therefore their section on clauses can be analysed to demonstrate LFS’s proficiency in producing subordinate clauses. It appears that many sentences were given to LFS in English, and she then translated them into Wappo. One type of clause that is not morphologically marked in Wappo is the reason clause, but LFS translated them using thuɁ, exemplifying the difficulties found when translating, and LFS’s ability to innovate (Thompson et.al 2006:114). An example given in the grammar is: cephi

o - paʔo - lahkhiʔ thuʔ č’oley - khiʔ

3SG:NOM UOP – eat – NEG

so skinny – STAT

‘s/he’s skinny because s/he doesn’t eat’ Thompson et.al (2006:129) state that there is some variation in LFS’s use of the causative, which may suggest that some forms may be lexicalised, or that LFS has not had stable access to infrequent causative verb forms.

32 2.5. Conclusion Overall, it appears that LFS had a good command of Wappo, even bringing Wappo calques into her English, which is arguably atypical of a speaker of a moribund language. Elmendorf stated that her Wappo was restricted to one level, but even so, she possessed a large vocabulary and an ability to create and translate complex sentences easily, even innovating for morphological markers that do not exist in Wappo. Sawyer (1991:14) commended LFS’s memory to recall Wappo vocabulary, and raised the point that although a speaker may seem like they are having lapses in knowledge, it may be due to the fact that English and indigenous American languages often do not translate well, causing difficulties for the consultant who is attempting to think of an adequate translation. She did however, have some self-confessed weaknesses. LFS’s zoological vocabulary, although perhaps one of her self-confessed weaker areas, was still not bad. Although there are some obvious gaps, the majority of the terms she produces are shared with Radin’s consultants. If zoological terms are considered to be one of the weaker areas of LFS’s vocabulary, this lends itself to suggest that she possessed a rather rich vocabulary overall.

33

3. The case of Tunica and its last fluent native speaker, Sesostrie Youchigant 3.1. The history and decline of Tunica The Tunica language is a language isolate which was spoken in the Lower Mississippi Valley by the Tunica tribe until the death of its last speaker Sesostrie Youchigant sometime after 1938 (Brain et.al 2004:586). There have been attempts by John Swanton in 1919 to group Tunica with Chitimacha and Atakapan as a ‘Tunican’ family, with Mary Haas later grouping these languages with Muskogean and Natchez to create a larger ‘Gulf’ family, some of which are highlighted in dark pink on Figure 6, but this has been rejected by linguists such as Campbell (Brain et.al 2004:586). The Tunica once occupied lands in Northern Mississippi and Arkansas around the 1540s, although they migrated numerous times throughout their history, often due to wars such as the 1706 war with the Chickasaw tribe Eventually, by 1771, the Tunica had migrated to their most recent residence, the area around Marksville, Louisiana. (Brain et.al 2004:586-9). By the end of the 19th century, the Tunica, Ofo, Avoyelles, Choctaw and Biloxi tribes lived in close proximity. As these tribes often socialised together, either Louisiana French or Mobilian Jargon was used, the latter of which was a pidgin language largely based on Choctaw and Chickasaw (Haas 1940:9). As a result of these tribes coming together and needing a common language to communicate, their own native languages fell into disuse.

34

Figure 6: Map of pre-contact Southeastern tribal languages, including Tunica.

Image taken from University of Louisiana at Lafayette, from Sturtevant (1967)

35 3.2. Sesostrie Youchigant’s (c.1870-a.1940-50) fluency in Tunica

Image taken from Ledbetter (2011) Tunica-Biloxi tribe honors past and builds future in the Washington Times.

36

Sesostrie Youchigant (SY) was born around 1870 in Marksville, Louisiana to Sosthene Youchigant and Arsene Chiki (Haas 1940:12, Downs 2009:83). SY could speak Louisiana French and English in addition to Tunica, although his French was better than his English. It may be likely that SY did not receive much schooling, as up until the 1940s, the Tunica were excluded from attending schools for white children, and refused to attend schools for black children (Bureau of Indian Affairs 1980:19). His mother would often attempt to converse with him in Tunica, but he would reply in French. SY was the chief of the Tunica tribe from 1911 until his resignation in 1921 (BoIA: 16). SY’s mother died in 1915, which left him with no other person to converse in Tunica with until Mary Haas conducted fieldwork with him in 1933-38, when he would have been around 63 years old (Downs 2009:82, Haas 1950:9). Haas conducted fieldwork with SY for three months in 1933, working out the grammatical structure of Tunica, and collecting texts. At that time, SY had not conversed in Tunica for around 18 years. Haas revisited SY four times between 1933 and 1938, and obtained more versions of the texts that she had obtained in 1933. The second and third versions of the myths SY told were usually more to his satisfaction, which indicates that throughout Haas’ visits, he had managed to recall more of the language than in 1933. Both SY and Haas acknowledged that he had gaps in his knowledge largely due to the long period of disuse of his language (Haas 1940:13). Haas believed that SY could repeat what he had heard, but he could not innovate with the language. Haas also added that lacunae in the material that she collected mostly corresponded to lacunae in SY’s knowledge. This may be a result of him choosing to interact with his mother in French rather than Tunica, a fact that he regretted in later life (Haas 1940:12). Haas in her 1953 Tunica Dictionary included both words given by SY, and also the Gatschet-Swanton vocabulary, which was collected separately by Albert Gatschet in 1886 with William Ely

37 Johnson, and by John Swanton in 1907-10, who checked Gatschet’s material with Johnson, but also obtained more data from SY’s maternal uncle, Volsine Chiki (Haas 1953:179). Swanton (1921:1) comments on Gatschet’s work, stating that there was not much more that he could add to the data that Gatschet collected from Johnson, aside from a few texts. SY however believed that Gatschet’s consultant Johnson had a better command of the Biloxi language than Tunica, but was a good interpreter of Louisiana French. Haas deemed Gatschet’s choice of consultant unfortunate, as at the time that he worked on Tunica, there were numerous other fluent speakers of the language (Haas 1953:179). By the time Swanton (1921:1) worked with Tunica, no more than a dozen people could speak it with anything approaching fluency, and no Tunica men had the ability to make good interpreters. 3.3. Knowledge of lexical items All of the roots in Haas’ Part I: Tunica-English dictionary have been analysed to discover the percentage of lexical items that Gatschet and Swanton collected that SY did not produce. Out of an estimated 1080 roots, 86 Gatschet-Swanton words were identified, which is around 7.9% of the entire dictionary. Of these 86 words, 54 were marked as words given solely by the Gatschet-Swanton vocabulary, while for the other 32, SY gave different words or definitions than Gatschet and Swanton’s informants. Of the 54 marked Gatschet-Swanton words, 30 of them were marked by Haas as ‘not known to Y’, which indicates that with the remaining 24 words, SY may have shown some recognition of their meanings, although he did not provide them himself. Three words in the GatschetSwanton vocabulary were identified by SY as words borrowed into Tunica from other languages, such as čula ‘fox’, being from Choctaw. There are also three words that SY recalled once, but later forgot, an example being núpali ‘to mend’. Similarly to the headwords, there are also affixes in the Gatschet-Swanton vocabulary. In total, 34 affixes have been identified as being recorded by Gatschet and Swanton, with

38 four marked as completely unknown to SY and seven having a different meaning given by SY. As a considerable amount of these affixes can be placed with numerous different roots, it is less likely that SY would be completely unaware of a large amount of affixes. Some affixes are only found in one word, such as –wúhi only occurring in kósuwúhi ‘rainbow’. The affixes that were completely unknown to SY were –hči ‘paternal aunt’, malia prefix which referred to birds, -niši ‘breast’ and –šari ‘to have time’. Haas notes that malimay have been an incorrect analysis by Gatschet, who frequently gave inaccurate bird terms, while for –niši, SY used ˀúču to refer to ‘breast’. A notable example of the lacunae in information in both the Gatschet-Swanton vocabulary and Haas’ vocabulary from SY is the incomplete list of Tunica month names given by all informants. Gatschet and Swanton’s informants named January through July, while SY named May through September, as well as an unidentified month name, táhčˀaméli (Haas 1953:264-5). 3.4. Tunica Texts In Tunica Texts, Haas collected various myths and stories from SY, accompanied by footnotes. There are two or three tellings of some texts, with some differences between them. Haas notes that SY does include some calques on English in his Tunica, such as ‘she set the table’ and ‘take the highway’ (Haas 1950:27, 93). In the first telling of the Tunica thunder myth, SY uses ‘his old woman’ as a calque on English, but in the second telling; he uses the correct Tunica word meaning ‘wife’ (Haas 1950:31). The flood myth given by SY had also been given by his maternal uncle Volsine Chiki to John Swanton, but both their versions are different, which Haas notes as odd due to the two consultants being closely related. Chiki concludes the flood myth with an incident SY doesn’t mention, while SY has two additional incidents that Chiki doesn’t mention. Haas believes that this may go back to a longer, more complete flood myth that Chiki and SY recall different fragments of.

39 Haas also notes that there is some possible Biblical influence in SY’s telling of the flood myth (Haas 1950:61). An interesting example of the evolution of SY’s Tunica from 1933-38 is given in the flood myth. SY uses the word ta’rusˀaha approximately ‘there is no knowing’ in the first telling. The verb ‘to know’, -e’rusa, is a static verb, therefore by the rules of Tunica grammar, it should be inflected by a personal pronominal prefix. Instead, it is inflected by the article ta’-. Haas believes that ta’rusˀaha is either an incorrect usage or an exception to the rule. It may be more likely to be the former, considering that in the second telling, SY did not use this word again (Haas 1950:61). 3.5. Conclusion From looking at SY’s language, particularly the examples given in Tunica Texts, his command of Tunica improves as he becomes more confident with his language. In the second and third tellings of some of the Tunica myths, SY provides some inadvertent corrections such as the calque ‘his old woman’ being replaced with the Tunica word for wife. Out of the three cases discussed in this paper, SY most likely had the longest period of continued disuse of his language, and does show more obvious lapses in memory, such as recalling some words once and then later forgetting them. Although there are noticeable gaps in SY’s knowledge, less than 10% of his vocabulary differs from that found in Gatschet and Swanton’s vocabulary. He therefore demonstrates at the very least that he has a vocabulary of about 1000 roots, although the size of his innate vocabulary may be higher. However, although there are lacunae in his vocabulary, he does possess the ability to tell myths in his language, with Haas describing his knowledge of Tunica mythology as much more balanced than anyone can expect it to be at that stage in the life of the Tunica language (Haas 1950:1).

40

Conclusion All three speakers studied in this paper demonstrate that they have a good command of their native language, albeit in different areas on different levels. Each of the consultants vary in who they learned their L1 from, their levels of education, their upbringings, and most significantly of all, the length of time in which they did not use their language at all. Both LFS and SY learned their native language from their mothers. NM however differed in that he first learned to speak English, and then learned Manx from his great aunt after going to live with her sometime between the ages of two and a half to five years old (Broderick 1999:75). Maddrell was immersed in the primarily Manx-speaking community of Cregneash from a young age, which, if child language acquisition theories of the critical period are to be supported, is within the period of childhood before lateralisation of the brain occurs around the onset of puberty (Chun 1980:288). Whether or not the critical period hypothesis is to be taken in consideration, NM’s Manx was nevertheless not overly influenced by English like many other speakers of LSM. NM and SY both had some form of education, with NM leaving school around the age of thirteen to work at sea. SY is able to read, but it is unclear as to how much if any formal education he had, considering that up until the 1940s, schools were segregated between white children and black and Indian children (Haas 1950:173). LFS, in contrast, did not receive any schooling at all, as she was a carer for her blind mother. As a result, LFS retained her Wappo into adulthood, before being exposed to English (Thompson et.al 2006:xiii). It seemed that contrary to many terminal speakers of a language, LFS’s English showed influence from Wappo, rather than the opposite. The most significant factor which affected how these three speakers used their language was the period of time in which they did not use it to communicate with any other native speaker.

41

Table 3: Lengths of time in which each speaker’s L1 was not used. NM 2.5-5 years

Age of acquisition of native language Approximate age No complete of beginning of disuse of language, language disuse but disuse of Manx as sole language of communication around age 13 Approximate age First record of NM when linguistic as a Manx research was informant age 72 in conducted 1950-1 by Jackson (1955:vi)

LFS From birth

SY From birth

Roughly age 78, after the death of her sister in the early 1970s (Thompson et.al 2006:xiii) At least age 64 sometime in the 1950s (Sawyer 1991:11)

Around age 45 after the death of his mother in 1915

Age 63 in 1933

Table 3 demonstrates the period of disuse of each speaker’s native language. It can be seen that neither NM nor LFS had an extensive period of disuse of the language. NM did likely have reduced exposure to Manx from around age thirteen, but not enough to greatly impact on his fluency of the language. LFS’s age when linguistic research was conducted actually overlaps with when she stopped using of the language for communication, which indicates that disuse of the language would not have been as significant a factor affecting her Wappo as it would have been for SY, who had the longest period of complete disuse of his language, not having spoken Tunica at all for eighteen years. In SY’s case, Haas (1950:9) notices that he improves his Tunica between her visits from 1933-38, suggesting a recalling process which LFS did not need to experience. NM also had to recall some of his Manx, as he stated that it almost left him after he went to sea, although when he returned to the Isle of Man, some, but not all of it returned to him (Broderick 1984:468). As all three speakers were over age sixty when they became consultants, it may be that their age could have caused lapses in their memory. However, the method of elicitation used by the researchers of these languages would have been crucial in obtaining as much

42 information about the language as possible. Each speaker is an individual, preferring different methods of elicitation. LFS did not like to give texts or stories in Wappo, but from looking at the work of Sawyer (1960-8), and of Thompson et.al (2006:114), the former tended to record word lists to elicit vocabulary from LFS, and the latter appeared to ask her to translate some sentences from English into Wappo. NM and SY did give a lot of texts in comparison to LFS. Jackson (1953:3) elicited lexical sets from his informants, but it was shown that NM was more comfortable when producing anecdotes than being asked to recall a set of lexemes. SY also demonstrated his command of Tunica better in Tunica Texts than when giving vocabulary. Dorian (1977:29-30) states that to assess the fluency of a speaker of a moribund language, comparisons to older forms of the language can determine how fluent the current speakers are, due to language reduction. All three speakers can be defined as terminal speakers, due to the aforementioned phenomenon. SY showed some loss of vocabulary since Gatschet and Swanton studied Tunica in the late 1800s and early 1900s, and although he produced many myths, most are fragmentary. This may be due to the decline in Tunica culture rather than a lapse in his memory, as both himself and Chiki managed to produce different fragments of the Tunica flood myth which did not always correspond with one another’s versions (Haas 1950:2, 61). LFS shows stylistic loss in her Wappo since Radin’s informant Tripo, who could use the Wappo style for telling myths (Elmendorf 1981:277). This could be attributed to either her gender, her lack of schooling which restricted her use of language registers, or both. NM shows surprisingly little reduction compared to other speakers of his time, but features such as the lack of distinction of gender in the definite article are seen in his speech. The question of how much of their native language that a terminal speaker knows can be answered by numerous different factors. It can be seen in the three case studies in this paper that each speaker is an individual, with different backgrounds. In order to assemble

43 as much as possible as to how much a terminal speaker knows of their language, the method of elicitation should be taken into consideration. LFS retained her Wappo much longer than SY or NM retained their languages due to her lack of an education. Perhaps the most significant factor in how much of their native language a terminal speaker will know is the length of time in which they did not use their language at all. SY’s situation is separate from NM and LFS’s as he did not use Tunica for eighteen years. However, without SY, only around 40% of the present lexical information of Tunica, and hardly any texts would be known. Similarly, without LFS, about a third less would be known about Wappo. In contrast, without NM, there would not be much Manx that would be unknown, however, NM is significant amongst other Manx speakers of his time in that his Manx gave a good picture of what Manx as a community language would have been like in the early 1800s, due learning the language from his great aunt. These case studies show that there is not and should not be one universal manner when determining how much a terminal speaker knows; each speaker prefers to communicate with their consultant in different ways, and some methods, such as lexical sets or the elicitation and study of texts may be more appropriate for some speakers, while less so for others.

44

Bibliography Abel-Vidor, S., Brovarney, D. & Billy, S. (1996) Remember Your Relations: The Elsie Allen Baskets, Family & Friends. Berkeley: Heyday. Baugh, A.C. & Cable, T. (1993) 5th ed. A History of the English Language. London: Routledge. Brain, J.P, Roth, G. & De Reuse, W.J. (2004) ‘Tunica, Biloxi and Ofo’ In: Fogelson, R. & Sturtevant, W. (eds.) Handbook of North American Indians, Volume 14: Southeast. Washington: Smithsonian Institution. pp.586-597. Broderick, G. (1984) A Handbook of Late Spoken Manx, Vol. 1: Grammar and Texts. Tübingen: Niemeyer. Broderick, G. (1993) ‘Manx’ In: Ball, M.J. & Fife, J. (eds.) The Celtic Languages. London: Routledge. pp.228-286 Broderick, G. (1999) Language Death in the Isle of Man: An investigation into the decline and extinction of Manx Gaelic as a community language in the Isle of Man. Tübingen: Niemeyer. Bureau of Indian Affairs (1980) Recommendation and summary of evidence for proposed finding for Federal acknowledgment of the Tunica-Biloxi Indian Tribe of Louisiana. Washington DC: United States Department of Interior. Campbell, L. & Muntzel, M.C. (1989) 'The structural consequences of language death' In: Dorian, N. (ed.) Investigating Obsolescence: Studies in language contraction and death. Cambridge: Cambridge University Press. pp.181-196. Chun, J. (1980) ‘A Survey of Research in Second Language Acquisition’ The Modern Language Journal. 64 (3) pp.287-296. Cregeen, A. (1835) A dictionary of the Manks language. Douglas. Dorian, N. (1977) ‘The Problem of the Semi-Speaker in Language Death’ International Journal of the Sociology of Language. 12 pp.23-32. Downs, E.C. (2009) ‘The struggle of the Louisiana Tunica Indians for Recognition’ In: Williams, W.L. (ed.) Southeastern Indians Since the Removal Era. Georgia: University of Georgia Press. pp.72-89. Driver, H.E. (1936) ‘Wappo Ethnography’. University of California Publications in American Archaeology and Ethnology. 36 (3) pp.179-220. Elmendorf, W.W. (1968) ‘Lexical and Cultural Change in Yukian’ Anthropological Linguistics. 10 (7) pp.1-41. Elmendorf, W.W. (1981) 'Last speakers and language change: two Californian cases'

45 Anthropological Linguistics. 23 (1) pp.36-49.

Golla, V. (2011) California Indian Languages. Berkeley, Los Angeles, London: University of California Press. Haas, M.R. (1940) Tunica. New York: J.J. Augustin Publisher. Haas, M.R. (1950) Tunica Texts. Berkeley, Los Angeles: University of California Press. Haas, M.R. (1953) Tunica Dictionary. Berkeley, Los Angeles: University of California Press. Hill, J. (1978) ‘Language Death, Language Contact and Language Evolution’. In: McCormack, W.C. & Wurm, S.A. (eds.) Approaches to Language: Anthropological Issues. The Hague: Mouton. pp.45-78. Jackson, K.H. (1955) Contributions to the study of Manx phonology. Edinburgh: Thomas Nelson. Jenner, H. (1875) ‘Manx language: Its grammar, literature and present state.’ The Philological Society, London. 18 June. Kelly, J. (1866) The Manx Dictionary. Douglas: The Manx Society. Kewley Draskau, J. (2004/5) ‘Language death and resurrection in the Isle of Man: the continuity of Manx Gaelic exemplified by the use of inflected verb tenses’. Proceedings of the Harvard Celtic Colloquium. 24/25 pp.229-246. Kewley Draskau, J. (2008) Practical Manx. Liverpool: Liverpool University Press. Mac Reachtain, L. (1993) ‘The World’s Oldest Gaelic Speaker’. In: Hughes, A.J. (ed.) Manx Speaker Ned Maddrell and Irish Gaelic author Liam Mac Reachtain. Seanchas Ardmhacha: Journal of the Armagh Diocesan Historical Society. 15 (2) pp.125-129. McGregor, W. (2009) Linguistics: An Introduction. London, New York: Continuum. Miller, S. (2007) ‘Here the Manx language lingers, and may linger some time longer: Manx and English in Cregneash in 1901’. Zeitschrift fur Celtische philologie. 55 pp.108-121 Quesada, J.D. (2000) 'Synopsis of a Borua Terminal Speaker'. Amérindia: Revue d'Ethnolinguistique Amérindienne. 25 pp.65-86. Radin, P. (1929) ‘A Grammar of the Wappo Language’ University of California Publications in American Archaeology and Ethnology. 27 pp.1-194. Sawyer, J.O. (1976) ‘Some Wappo Names for People and Languages’ The Journal of California Anthropology 3 (1) pp.120-127. Sawyer, J.O. (1991) Wappo Studies. (ed.) Shepherd, A. Berkeley: Survey of California and Other Indian Languages (Report 7).

46 Shepherd, A. (1991) ‘Foreword’ In: Sawyer, J.O. Wappo Studies. Berkeley: Survey of California and Other Indian Languages (Report 7). Stowell, B. (2005) ‘The case of Manx Gaelic/Yn Ghailckagh’ In: O’Neill, D. (ed.) Rebuilding the Celtic languages: Reversing language shift in the Celtic countries. Tal-y-bont: Y Lolfa. pp.383-416. Swanton, J.R. (1921) ‘The Tunica Language’ International Journal of American Linguistics 2 (1/2) pp.1-39. Tadmor, U. (2009) ‘Loanwords in the world’s languages: Findings and results’. In: Haspelmath, M. & Tadmor, U. (eds.) Loanwords in the World’s Languages: A Comparative Handbook. pp.55-75. Berlin: De Gruyter. Thompson, S.A., Park, J.S. & Li, C.N. (2006) A Reference Grammar of Wappo. Berkeley, Los Angeles, London: University of California Press. Tsitsipis, L. (1989) 'Skewed performance and full performance in language obsolecence: The case of an Albanian variety' In: Dorian, N. (ed.) Investigating Obsolecence: Studies in language contraction and death. Cambridge: Cambridge University Press. pp. 117-137. Voegelin, C.F. & Voegelin, F.M. (1977) ‘Is Tübatulabal De-Acquisition Relevant to Theories of Language Acquisition?’ International Journal of American Linguistics. 43 (4) pp.333-338.

Web Links Ballaugh Heritage Trust (2008) ‘John Kneen (1852-1958)’. http://ballaugh.org/2008/10/john-kneen-1852-1958 [accessed 16 December 2013] Dry Creek Rancheria (2014) Portrait of a Dry Creek Legend: Laura Fish Somersal. http://drycreekrancheria.com/portrait-of-a-dry-creek-legend-laura-somersal/ [accessed 11 March 2014] Fennig, C. (2013) 'Nailing Jello to the wall: Language Identification'. Ethnoblog. Ethnologue http://www.ethnologue.com/ethnoblog/chuck-fennig/nailing-jello-wall-languageidentification [accessed 23 October 2013] Isle of Man Family History Society (2013) Burials. http://www.iomfhs.im/lawsons/LawsonBMD/burials/tx_taubman.html [accessed 2 October 2013] Ledbetter, C. (2011) Tunica-Biloxi tribe honors past and builds future. The Washington Times http://communities.washingtontimes.com/neighborhood/carlaledbetter/2011/feb/5/tunica-biloxi-tribe-honors-past-and-builds-future/ [accessed 21 April 2014] Lewis, Paul, M., Simons, G.F., Fennig, C.D. (2013) Ethnologue: Languages of the World, Seventeenth Edition. Dallas: SIL International. http://www.ethnologue.com/

47 Mason, B.J. (n.d.) ‘California Indians Peoples’ http://kacha.com/California%20Indians%20IAIR%20article%20pg1.htm [accessed 1 May 2014] Merriam, C.H. (c.1898) ‘Mi-Yahk-Mah (Southern Division of Wappo): English-Indian Dictionary’. archive.org (2014) C. Hart Merriam papers relating to work with California Indians, 1850-1974. https://archive.org/details/bancroft_chartmerriam_1556_62 [accessed February 22 2014] Miller, S. (1993) ‘The Death of Manx from newspaper clipping 1950s’. Gaelic-L Archives, LISTSERV. https://listserv.heanet.ie/ [accessed 16 November 2013] Moore, A.W. (1889) ‘Goldie-Taubmans of the Nunnery’ In: Manx Families. http://www.isleof-man.com/manxnotebook/famhist/families/awm1889/goldie.htm [accessed 16 December 2013] Ned Maddrell (n.d.) photograph http://www.flickr.com/photos/greinneyder/5486488481/ [accessed 24 November 2013] Sawyer, J.O. (1959-60) Unpublished Wappo field notes. California Language Archive: Image 80. http://cla.berkeley.edu/item/2836?tab=digital [accessed 20 February 2014] Sawyer, J.O. (1960-8) ‘Wappo audio recordings (1960-1968)’ California Language Archive http://cla.berkeley.edu/collection/10057 [accessed 25 April 2014] Sturtevant, W.C. (1967) ‘Early Indian tribes, culture areas, and linguistic stocks’. University of Lousiana at Lafayette http://www.ucs.louisiana.edu/~ras2777/indianlaw/indianlanguages.jpg [accessed 1 May 2014] Wood, G.W. (1896) ‘An account of the translation and editions of the holy scriptures in the Manx language.’ In: The Manx Church Magazine Vol.6. http://www.isle-ofman.com/manxnotebook/history/manks/script.htm [accessed 13 December 2013]