Problems of phonetic transcription: The case of the Hiberno-English slit-t

11
PHONETIC REPRESENTATION Problems of phonetic transcription: The case of the Hiberno-English slit-t HELEN PANDELI School of Environmental and Human Sciences, University of Wales Institute Cardiff, Western Avenue, Cardiff CF5 2YB. Wales e-mail: [email protected] JOSEPH F. ESKA Department of English, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0112. USA e-mail: [email protected] MARTIN J. BALL School of Behavioural and Communication Sciences, University of Ulster at Jordanstown, Newtownabbey, Co. Antrim BT37 OQB. N. Ireland e-mail: [email protected] AND JOAN RAHILLY School of English, Queen's University of Belfast, Belfast, BT7 INN. N. Ireland e-mail: [email protected] Acoustic and electropalatographic data on the so-called Hiberno-English 'slit- t' are reported, and the implications these data have for an adequate transcription are discussed. Previous transcription suggestions highlight the difficulty posed by the lack of an IPA diacritic for tongue shape. We conclude that the adoption of an alveolar diacritic (as used in the extensions to the IPA for transcribing disordered speech) could get round these difficulties. 1. Phonetic Symbolisation It is widely accepted that phonetic description, whether based on impressionistic transcription or instrumental analysis, can never be entirely accurate. This is because the physiological and acoustic properties of speech are far too complex for us to describe every detail. We are forced to make decisions about which properties we will describe and which we will ignore (aside from properties which we may be unable to record). Furthermore, every time we make a phonetic transcription in the form of a string of symbols suggesting discrete segments, we are compromising what we know about how speech is produced (see e.g. Ladefoged, 1990 on what we transcribe, and IPA 1995 on segmentation). Nevertheless we clearly need to be as accurate as possible when making phonetic, and in particular narrow phonetic, transcriptions of speech. With phonological transcriptions, where an analysis of only the contrastive sound units of a language or dialect is required, precision in the choice of symbols is not paramount. With narrow phonetic transcriptions, however, careful choice of symbols, and absence of ambiguity in what those symbols represent is all important. Journal of the International Phonetic Association 27: 65-75 at http:/www.cambridge.org/core/terms. http://dx.doi.org/10.1017/S0025100300005430 Downloaded from http:/www.cambridge.org/core. University Libraries - Virginia Tech, on 13 Oct 2016 at 14:56:08, subject to the Cambridge Core terms of use, available

Transcript of Problems of phonetic transcription: The case of the Hiberno-English slit-t

PHONETIC REPRESENTATION

Problems of phonetic transcriptionThe case of the Hiberno-English slit-t

HELEN PANDELISchool of Environmental and Human Sciences University of Wales Institute Cardiff

Western Avenue Cardiff CF5 2YB Walese-mail hpandeliuwicacuk

JOSEPH F ESKADepartment of English Virginia Polytechnic Institute and State University

Blacksburg Virginia 24061-0112 USAe-mail eskavtaixccvtedu

MARTIN J BALLSchool of Behavioural and Communication Sciences University of Ulster at Jordanstown

Newtownabbey Co Antrim BT37 OQB N Irelande-mail mjballulstacuk

AND JOAN RAHILLYSchool of English Queens University of Belfast

Belfast BT7 INN N Irelande-mail jrahillyclioartsqubacuk

Acoustic and electropalatographic data on the so-called Hiberno-English slit-t are reported and the implications these data have for an adequatetranscription are discussed Previous transcription suggestions highlight thedifficulty posed by the lack of an IPA diacritic for tongue shape We concludethat the adoption of an alveolar diacritic (as used in the extensions to the IPAfor transcribing disordered speech) could get round these difficulties

1 Phonetic Symbolisation

It is widely accepted that phonetic description whether based on impressionistictranscription or instrumental analysis can never be entirely accurate This is because thephysiological and acoustic properties of speech are far too complex for us to describeevery detail We are forced to make decisions about which properties we will describe andwhich we will ignore (aside from properties which we may be unable to record)Furthermore every time we make a phonetic transcription in the form of a string ofsymbols suggesting discrete segments we are compromising what we know about howspeech is produced (see eg Ladefoged 1990 on what we transcribe and IPA 1995 onsegmentation)

Nevertheless we clearly need to be as accurate as possible when making phonetic andin particular narrow phonetic transcriptions of speech With phonological transcriptionswhere an analysis of only the contrastive sound units of a language or dialect is requiredprecision in the choice of symbols is not paramount With narrow phonetic transcriptionshowever careful choice of symbols and absence of ambiguity in what those symbolsrepresent is all important

Journal of the International Phonetic Association 27 65-75

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

66 PANDELI ESKA BALL amp RAHILLY

Despite a widespread understanding that many different parameters are controlled inthe articulation of speech the base symbols of the International Phonetic Alphabet giveinformation about only a small number of the parameters For consonants for instancethe majority of symbols indicate explicitly only the place of articulation manner ofarticulation and voicing characteristics of the sounds they represent Diacritics can beadded to the base symbols to give more detail about for instance the identity of the activearticulator but there are still a number of features of the articulation of speech soundswhich cannot be transcribed One of these is cross-sectional tongue shape for lingualfricatives articulated in the anterior part of the oral cavity It has been shown that thesefricatives may differ considerably from one another in terms of this parameter (eg Stoneet al 1992) with some fricatives for instance exhibiting a flat or slit tongue shapeand others a grooved shape There is however currently no accepted IPA way oftranscribing this difference In this paper we deal with a transcription problem in this areanamely the transcription of a sound which occurs in southern Hiberno-English which weshall refer to as slit-t

2 The Occurrence of Hiberno-English Slit-t

The phenomenon of the slit-t (or ll lenition) is widespread in all varieties of southernHiberno-English Indeed Hickey (1986 17) says that in present day IrE this is themost widespread feature in the entire Republic apart from post-vocalicpre-consonantal hiand syllable final alveolar [= clear] I There is as yet no evidence to suggest that itoccurs in other varieties Hickey (1996 224) refers to the slit-t as particularly Irish Hemakes the case for the particular salience of the slit-t in southern Hiberno-English when hesays that t-lenition would appear to be confined to southern Irish English and not to befound in either mainland English or extra-territorial varieties of the language (p 225)While there is some anecdotal evidence to suggest that t-lenition is becoming morecommon in some northern Hiberno-English varieties it remains true that the feature isparticularly characteristic of the south

In southern Hiberno-English the slit-t realisation occurs intervocalically (eg as insitting) and word-finally (as in but for example) However if another consonantimmediately precedes or follows the l] lenition will not occur (eg backed bootleg)Lenition may also occur with the voiced apico-alveolar fricative as in fodder The nature ofthis form has not been fully explored but may be a slit-d

It has been suggested (Wells 1982) that the presence of the slit-t is due to interferencefrom Irish but this remains to be demonstrated Hickey (1984b) suggests that theweakening of Ixl in southern Hiberno-English is a modern parallel to the system of coronalobstruents in Old High German where a contrast between an apico-alveolar fricative and alamino-alveolar fricative existed

3 The Physical Characteristics of Slit-t

In spite of the salience of the slit-t in southern Hiberno-English dialects detailedarticulatory descriptions are rare There are however various terms available which seemto capture the same sound Nf Ghallchoir (1981 163) in the case of north-west Donegal

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 67

for example refers to an allophonic dental plosive accompanied by fricativeaspiration as in throaf Tilling (1981 100) for Kinlough Co Leitrim also states that IMin final position is generally realised with aspiration or with affrication and notes thatIxl without either aspiration or affrication are occasional in this variety

A more useful articulatory description occurs in Hickey (1996) who refers to thesound as an apico-alveolar fricative He points out that this sound requires the tip of thetongue as the active articulator making it thus distinctive from the fricatives s and ITJ butis therefore different from bus and buzz- Further comment on the articulation of the soundis provided by Hickey in the following descriptions

(1) It is formed by bringing the apex of the tongue close to the alveolar ridge as if forthe articulation of ll but stopping just before contact (Hickey 1984a 234)

[It] is a controlled articulatory gesture which moves the apex of the tongue towardthe alveolar ridge and holds it there for a duration approximately that of otherfricatives It should be stressed here that [slit-t] never involves apico-alveolarclosure however brief and that it is a controlled movement which is considerablylonger than a tap (Hickey 1984b 676)

[It] has frequently been referred to as a flap but in fact it is a controlled articulationwhich results from the tongue being held just below the alveolar ridge to producean apico-alveolar fricative rather than a stop (Hickey 1995124)

A more detailed picture of the articulatory characteristics of slit-t can be gained fromEPG data presented in Pandeli (1993) The data were collected from a male speaker fromDublin in his mid twenties 9 tokens each of the words neat met mat and put wererecorded They were read in a list alongside words containing other fricatives using thecarrier phrase have a This gave a total of 36 word-final tokens of slit-t in fourdifferent vowel environments

For each token the pattern of contact between the tongue and the palate was examinedduring a single target EPG frame in which the number of activated electrodes in theregion of the major constriction was at a maximum The pattern of contact during the target

slit-t [s] [fl

Figure 1 Linguopalatal contact patterns during a single target EPG framefor slit-t [s] and [f] in met mess and mesh respectively

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

68 PANDELI ESKA BALL amp RAHILLY

frame of one typical token of the fricative in the word met is shown in Figure 1

Analysis of the data revealed that contact was concentrated at the sides of the palateacross an area stretching from the front to the middle or back of the alveolar ridge A ratherbroad fricative channel was evident whose mean width was estimated to be 104 mmThis width was very similar to that estimated for tokens of [J] produced by the samespeaker (102 mm) but significantly greater than that for tokens of [s] (60 mm) TypicalEPG contact patterns for [s] and [J] are also shown in Figure 1 Finally the anterior edgeof the area of contact between the tongue and the palate was typically rather flat in shapefor slit-t with no contact at the sides of the palate in front of the major constrictionPandeli (1993) argues that this type of contact pattern is indicative of an apical articulation

In summary then the articulatory data reported in Pandeli (1993) suggest that for thisspeaker at least slit-t is an apical alveolar fricative formed with a broad central channelThe tongue shape which produces this type of fricative channel has been described in theliterature as eg flat (Pike 1943) and slit (Hockett 1955 Ladefoged 1971)

i IliUllMIIIlitilliiiitttiiilllll0231 Tine (sec) 0740

Figure 2 Spectrogram of met

In order to gain an impression of the durational characteristics of slit-t spectrogramswere made from acoustic recordings made at the same time as the EPG recordings Anexample spectrogram of the word met is shown in Figure 2 The duration of each of the 36tokens of slit-t was measured on the corresponding spectrogram and mean durationscalculated of the fricative in each of the four test words and in all words taken togetherFor comparison measurements were also made of the duration of [s] in 9 tokens each ofthe words lease mess mass and puss and of [J] in 9 tokens of leash mesh mash andpush These words were produced by the same speaker and recorded during the samesession as the slit-t words Mean durations of the three fricatives are shown in Table 1 Itcan be seen that the duration of all three fricatives is rather long reflecting their word-finalutterance-final position as well as a rather slow speaking rate Slit-t has a somewhatshorter mean duration (239 ms) than [s] (290 ms) and [f] (259 ms)

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 69

Table 1 Mean duration (ms) of Hiberno-English fricatives (9 tokens of each word)

slit-t

S

J

Neat

246

Lease

296

Leash

276

Met

245

Mess

284

Mesh

261

Mat

221

Mass

272

Mash

238

Put

243

306

Push

261

All

239

AH

290

All

259

4 Symbolising Slit-t

Given the physical characteristics of slit-t outlined above we will now consider howthe sound might best be transcribed The IPA lacks any specific symbol for the soundand more generally lacks any diacritic to indicate cross-sectional tongue shape (whetherflat or grooved) As a result a plethora of ad hoc transcriptions of slit-t can be found in theliterature Those known to us are the following

(2) [r] (Harris 1984 139 Lass 1987 267)

[s]

[]

[]

[s]

[si

[ri

mft]ft]

ft]

ft]

[ts]

[6]

(Harris 1990 285)

(Bertz 1987 45)

(Bertz 1987 45)

(Bertz 1987 45)

(Pilch 1959 420)

(Hogan 1934 18 Henry

(Conrick 198173)

(6Baoill 1990 161)1

(Wells 1982 429 Kallen

(IPA 1995 15)

(Hickey 1984a 234-235

(Henry 1958 123 Barry

(Henry 1958 123)

(Laver 1994 260)2

1958 123)

1994178)

1984b 675-676 1993 221-222 1995 123-124)

198168)

1 We have been informed by 6 Baoill (personal communication 25 ix 1995) that this perplexingtranscription mdash how can [t] be raised mdash is not his but was suggested by an editor of the volume in whichhis paper appears owing to the difficulty of setting his preferred transcription viz [ t] typographically

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

70 PANDELI ESKA BALL amp RAHILLY

Some of these transcriptions are simply idiosyncratic others are misleading or indeedinaccurate It can be seen from the above list that transcribers have generally used as basesymbols the following [r] [s] [t] and [8] We will now look at each of these in turn

[r]

This symbol represents a tap (or flap or flick in some terminology) which isessentially a very short sound Catford (1977129-30) states that the hit registered by aflick or by a passing transient flap has a measurable duration of the order of 1 to 3centiseconds and Ladefoged amp Maddieson (1996) cite Quilis (1981) who reports that themean closure duration of tokens of Spanish [r] which he measured was 20 ms As themean duration of Hiberno-English slit-t in the data presented above is 239 ms this clearlyrules out any consideration of the sound as a tap flap It would also be expected that analveolar tap would normally be formed with a complete closure on the alveolar ridge(though Laver 1994 reports the existence of tapped fricatives) Such a closure is notfound for slit-t [r] must therefore be discounted as the base symbol for this sound

[s]

The choice of [s] as a symbol is presumably motivated by an attempt to demonstratethe fricative nature of slit-t as well as its alveolar place of articulation We feel howeverthat this symbol is bound to be misleading as the IPA does not have any recogniseddiacritics to indicate the cross-sectional shape of the tongue This is a problem as the datapresented above suggest that the tongue shape for slit-t is flat (as for []) whereas for [s] itis grooved

[t]

With [t] the problem of cross-sectional tongue shape is less important The basesymbol [t] together with the lowering diacritic (ie [t ] = []) suggests a fricative mannerof articulation but not the grooved tongue shape of [s]

However the symbol [t] is ambiguous in terms of place of articulation On the onehand the Principles of the IPA (IPA 1949 17) state When it is desired to show inwriting that t d n are alveolar and not dental the retraction sign in the form - may beplaced underneath thus t d n This convention is still being adhered to by a number ofresearchers in the transcription of languages which contrast dental and alveolarconsonants For example Spajic et al (1996) use the retraction diacritic to transcribealveolar consonants in Toda (contrasting with a series of dental consonants) This sameexample is quoted in Ladefoged and Maddieson (1996 22) where the voiceless alveolarplosive of Toda is transcribed as [t]

2 Ladefoged and Maddieson (1996 144) also employ [8] to transcribe a voiceless flat alveolar fricative butdo not cite examples from southern Hiberno-English

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

66 PANDELI ESKA BALL amp RAHILLY

Despite a widespread understanding that many different parameters are controlled inthe articulation of speech the base symbols of the International Phonetic Alphabet giveinformation about only a small number of the parameters For consonants for instancethe majority of symbols indicate explicitly only the place of articulation manner ofarticulation and voicing characteristics of the sounds they represent Diacritics can beadded to the base symbols to give more detail about for instance the identity of the activearticulator but there are still a number of features of the articulation of speech soundswhich cannot be transcribed One of these is cross-sectional tongue shape for lingualfricatives articulated in the anterior part of the oral cavity It has been shown that thesefricatives may differ considerably from one another in terms of this parameter (eg Stoneet al 1992) with some fricatives for instance exhibiting a flat or slit tongue shapeand others a grooved shape There is however currently no accepted IPA way oftranscribing this difference In this paper we deal with a transcription problem in this areanamely the transcription of a sound which occurs in southern Hiberno-English which weshall refer to as slit-t

2 The Occurrence of Hiberno-English Slit-t

The phenomenon of the slit-t (or ll lenition) is widespread in all varieties of southernHiberno-English Indeed Hickey (1986 17) says that in present day IrE this is themost widespread feature in the entire Republic apart from post-vocalicpre-consonantal hiand syllable final alveolar [= clear] I There is as yet no evidence to suggest that itoccurs in other varieties Hickey (1996 224) refers to the slit-t as particularly Irish Hemakes the case for the particular salience of the slit-t in southern Hiberno-English when hesays that t-lenition would appear to be confined to southern Irish English and not to befound in either mainland English or extra-territorial varieties of the language (p 225)While there is some anecdotal evidence to suggest that t-lenition is becoming morecommon in some northern Hiberno-English varieties it remains true that the feature isparticularly characteristic of the south

In southern Hiberno-English the slit-t realisation occurs intervocalically (eg as insitting) and word-finally (as in but for example) However if another consonantimmediately precedes or follows the l] lenition will not occur (eg backed bootleg)Lenition may also occur with the voiced apico-alveolar fricative as in fodder The nature ofthis form has not been fully explored but may be a slit-d

It has been suggested (Wells 1982) that the presence of the slit-t is due to interferencefrom Irish but this remains to be demonstrated Hickey (1984b) suggests that theweakening of Ixl in southern Hiberno-English is a modern parallel to the system of coronalobstruents in Old High German where a contrast between an apico-alveolar fricative and alamino-alveolar fricative existed

3 The Physical Characteristics of Slit-t

In spite of the salience of the slit-t in southern Hiberno-English dialects detailedarticulatory descriptions are rare There are however various terms available which seemto capture the same sound Nf Ghallchoir (1981 163) in the case of north-west Donegal

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 67

for example refers to an allophonic dental plosive accompanied by fricativeaspiration as in throaf Tilling (1981 100) for Kinlough Co Leitrim also states that IMin final position is generally realised with aspiration or with affrication and notes thatIxl without either aspiration or affrication are occasional in this variety

A more useful articulatory description occurs in Hickey (1996) who refers to thesound as an apico-alveolar fricative He points out that this sound requires the tip of thetongue as the active articulator making it thus distinctive from the fricatives s and ITJ butis therefore different from bus and buzz- Further comment on the articulation of the soundis provided by Hickey in the following descriptions

(1) It is formed by bringing the apex of the tongue close to the alveolar ridge as if forthe articulation of ll but stopping just before contact (Hickey 1984a 234)

[It] is a controlled articulatory gesture which moves the apex of the tongue towardthe alveolar ridge and holds it there for a duration approximately that of otherfricatives It should be stressed here that [slit-t] never involves apico-alveolarclosure however brief and that it is a controlled movement which is considerablylonger than a tap (Hickey 1984b 676)

[It] has frequently been referred to as a flap but in fact it is a controlled articulationwhich results from the tongue being held just below the alveolar ridge to producean apico-alveolar fricative rather than a stop (Hickey 1995124)

A more detailed picture of the articulatory characteristics of slit-t can be gained fromEPG data presented in Pandeli (1993) The data were collected from a male speaker fromDublin in his mid twenties 9 tokens each of the words neat met mat and put wererecorded They were read in a list alongside words containing other fricatives using thecarrier phrase have a This gave a total of 36 word-final tokens of slit-t in fourdifferent vowel environments

For each token the pattern of contact between the tongue and the palate was examinedduring a single target EPG frame in which the number of activated electrodes in theregion of the major constriction was at a maximum The pattern of contact during the target

slit-t [s] [fl

Figure 1 Linguopalatal contact patterns during a single target EPG framefor slit-t [s] and [f] in met mess and mesh respectively

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

68 PANDELI ESKA BALL amp RAHILLY

frame of one typical token of the fricative in the word met is shown in Figure 1

Analysis of the data revealed that contact was concentrated at the sides of the palateacross an area stretching from the front to the middle or back of the alveolar ridge A ratherbroad fricative channel was evident whose mean width was estimated to be 104 mmThis width was very similar to that estimated for tokens of [J] produced by the samespeaker (102 mm) but significantly greater than that for tokens of [s] (60 mm) TypicalEPG contact patterns for [s] and [J] are also shown in Figure 1 Finally the anterior edgeof the area of contact between the tongue and the palate was typically rather flat in shapefor slit-t with no contact at the sides of the palate in front of the major constrictionPandeli (1993) argues that this type of contact pattern is indicative of an apical articulation

In summary then the articulatory data reported in Pandeli (1993) suggest that for thisspeaker at least slit-t is an apical alveolar fricative formed with a broad central channelThe tongue shape which produces this type of fricative channel has been described in theliterature as eg flat (Pike 1943) and slit (Hockett 1955 Ladefoged 1971)

i IliUllMIIIlitilliiiitttiiilllll0231 Tine (sec) 0740

Figure 2 Spectrogram of met

In order to gain an impression of the durational characteristics of slit-t spectrogramswere made from acoustic recordings made at the same time as the EPG recordings Anexample spectrogram of the word met is shown in Figure 2 The duration of each of the 36tokens of slit-t was measured on the corresponding spectrogram and mean durationscalculated of the fricative in each of the four test words and in all words taken togetherFor comparison measurements were also made of the duration of [s] in 9 tokens each ofthe words lease mess mass and puss and of [J] in 9 tokens of leash mesh mash andpush These words were produced by the same speaker and recorded during the samesession as the slit-t words Mean durations of the three fricatives are shown in Table 1 Itcan be seen that the duration of all three fricatives is rather long reflecting their word-finalutterance-final position as well as a rather slow speaking rate Slit-t has a somewhatshorter mean duration (239 ms) than [s] (290 ms) and [f] (259 ms)

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 69

Table 1 Mean duration (ms) of Hiberno-English fricatives (9 tokens of each word)

slit-t

S

J

Neat

246

Lease

296

Leash

276

Met

245

Mess

284

Mesh

261

Mat

221

Mass

272

Mash

238

Put

243

306

Push

261

All

239

AH

290

All

259

4 Symbolising Slit-t

Given the physical characteristics of slit-t outlined above we will now consider howthe sound might best be transcribed The IPA lacks any specific symbol for the soundand more generally lacks any diacritic to indicate cross-sectional tongue shape (whetherflat or grooved) As a result a plethora of ad hoc transcriptions of slit-t can be found in theliterature Those known to us are the following

(2) [r] (Harris 1984 139 Lass 1987 267)

[s]

[]

[]

[s]

[si

[ri

mft]ft]

ft]

ft]

[ts]

[6]

(Harris 1990 285)

(Bertz 1987 45)

(Bertz 1987 45)

(Bertz 1987 45)

(Pilch 1959 420)

(Hogan 1934 18 Henry

(Conrick 198173)

(6Baoill 1990 161)1

(Wells 1982 429 Kallen

(IPA 1995 15)

(Hickey 1984a 234-235

(Henry 1958 123 Barry

(Henry 1958 123)

(Laver 1994 260)2

1958 123)

1994178)

1984b 675-676 1993 221-222 1995 123-124)

198168)

1 We have been informed by 6 Baoill (personal communication 25 ix 1995) that this perplexingtranscription mdash how can [t] be raised mdash is not his but was suggested by an editor of the volume in whichhis paper appears owing to the difficulty of setting his preferred transcription viz [ t] typographically

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

70 PANDELI ESKA BALL amp RAHILLY

Some of these transcriptions are simply idiosyncratic others are misleading or indeedinaccurate It can be seen from the above list that transcribers have generally used as basesymbols the following [r] [s] [t] and [8] We will now look at each of these in turn

[r]

This symbol represents a tap (or flap or flick in some terminology) which isessentially a very short sound Catford (1977129-30) states that the hit registered by aflick or by a passing transient flap has a measurable duration of the order of 1 to 3centiseconds and Ladefoged amp Maddieson (1996) cite Quilis (1981) who reports that themean closure duration of tokens of Spanish [r] which he measured was 20 ms As themean duration of Hiberno-English slit-t in the data presented above is 239 ms this clearlyrules out any consideration of the sound as a tap flap It would also be expected that analveolar tap would normally be formed with a complete closure on the alveolar ridge(though Laver 1994 reports the existence of tapped fricatives) Such a closure is notfound for slit-t [r] must therefore be discounted as the base symbol for this sound

[s]

The choice of [s] as a symbol is presumably motivated by an attempt to demonstratethe fricative nature of slit-t as well as its alveolar place of articulation We feel howeverthat this symbol is bound to be misleading as the IPA does not have any recogniseddiacritics to indicate the cross-sectional shape of the tongue This is a problem as the datapresented above suggest that the tongue shape for slit-t is flat (as for []) whereas for [s] itis grooved

[t]

With [t] the problem of cross-sectional tongue shape is less important The basesymbol [t] together with the lowering diacritic (ie [t ] = []) suggests a fricative mannerof articulation but not the grooved tongue shape of [s]

However the symbol [t] is ambiguous in terms of place of articulation On the onehand the Principles of the IPA (IPA 1949 17) state When it is desired to show inwriting that t d n are alveolar and not dental the retraction sign in the form - may beplaced underneath thus t d n This convention is still being adhered to by a number ofresearchers in the transcription of languages which contrast dental and alveolarconsonants For example Spajic et al (1996) use the retraction diacritic to transcribealveolar consonants in Toda (contrasting with a series of dental consonants) This sameexample is quoted in Ladefoged and Maddieson (1996 22) where the voiceless alveolarplosive of Toda is transcribed as [t]

2 Ladefoged and Maddieson (1996 144) also employ [8] to transcribe a voiceless flat alveolar fricative butdo not cite examples from southern Hiberno-English

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 67

for example refers to an allophonic dental plosive accompanied by fricativeaspiration as in throaf Tilling (1981 100) for Kinlough Co Leitrim also states that IMin final position is generally realised with aspiration or with affrication and notes thatIxl without either aspiration or affrication are occasional in this variety

A more useful articulatory description occurs in Hickey (1996) who refers to thesound as an apico-alveolar fricative He points out that this sound requires the tip of thetongue as the active articulator making it thus distinctive from the fricatives s and ITJ butis therefore different from bus and buzz- Further comment on the articulation of the soundis provided by Hickey in the following descriptions

(1) It is formed by bringing the apex of the tongue close to the alveolar ridge as if forthe articulation of ll but stopping just before contact (Hickey 1984a 234)

[It] is a controlled articulatory gesture which moves the apex of the tongue towardthe alveolar ridge and holds it there for a duration approximately that of otherfricatives It should be stressed here that [slit-t] never involves apico-alveolarclosure however brief and that it is a controlled movement which is considerablylonger than a tap (Hickey 1984b 676)

[It] has frequently been referred to as a flap but in fact it is a controlled articulationwhich results from the tongue being held just below the alveolar ridge to producean apico-alveolar fricative rather than a stop (Hickey 1995124)

A more detailed picture of the articulatory characteristics of slit-t can be gained fromEPG data presented in Pandeli (1993) The data were collected from a male speaker fromDublin in his mid twenties 9 tokens each of the words neat met mat and put wererecorded They were read in a list alongside words containing other fricatives using thecarrier phrase have a This gave a total of 36 word-final tokens of slit-t in fourdifferent vowel environments

For each token the pattern of contact between the tongue and the palate was examinedduring a single target EPG frame in which the number of activated electrodes in theregion of the major constriction was at a maximum The pattern of contact during the target

slit-t [s] [fl

Figure 1 Linguopalatal contact patterns during a single target EPG framefor slit-t [s] and [f] in met mess and mesh respectively

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

68 PANDELI ESKA BALL amp RAHILLY

frame of one typical token of the fricative in the word met is shown in Figure 1

Analysis of the data revealed that contact was concentrated at the sides of the palateacross an area stretching from the front to the middle or back of the alveolar ridge A ratherbroad fricative channel was evident whose mean width was estimated to be 104 mmThis width was very similar to that estimated for tokens of [J] produced by the samespeaker (102 mm) but significantly greater than that for tokens of [s] (60 mm) TypicalEPG contact patterns for [s] and [J] are also shown in Figure 1 Finally the anterior edgeof the area of contact between the tongue and the palate was typically rather flat in shapefor slit-t with no contact at the sides of the palate in front of the major constrictionPandeli (1993) argues that this type of contact pattern is indicative of an apical articulation

In summary then the articulatory data reported in Pandeli (1993) suggest that for thisspeaker at least slit-t is an apical alveolar fricative formed with a broad central channelThe tongue shape which produces this type of fricative channel has been described in theliterature as eg flat (Pike 1943) and slit (Hockett 1955 Ladefoged 1971)

i IliUllMIIIlitilliiiitttiiilllll0231 Tine (sec) 0740

Figure 2 Spectrogram of met

In order to gain an impression of the durational characteristics of slit-t spectrogramswere made from acoustic recordings made at the same time as the EPG recordings Anexample spectrogram of the word met is shown in Figure 2 The duration of each of the 36tokens of slit-t was measured on the corresponding spectrogram and mean durationscalculated of the fricative in each of the four test words and in all words taken togetherFor comparison measurements were also made of the duration of [s] in 9 tokens each ofthe words lease mess mass and puss and of [J] in 9 tokens of leash mesh mash andpush These words were produced by the same speaker and recorded during the samesession as the slit-t words Mean durations of the three fricatives are shown in Table 1 Itcan be seen that the duration of all three fricatives is rather long reflecting their word-finalutterance-final position as well as a rather slow speaking rate Slit-t has a somewhatshorter mean duration (239 ms) than [s] (290 ms) and [f] (259 ms)

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 69

Table 1 Mean duration (ms) of Hiberno-English fricatives (9 tokens of each word)

slit-t

S

J

Neat

246

Lease

296

Leash

276

Met

245

Mess

284

Mesh

261

Mat

221

Mass

272

Mash

238

Put

243

306

Push

261

All

239

AH

290

All

259

4 Symbolising Slit-t

Given the physical characteristics of slit-t outlined above we will now consider howthe sound might best be transcribed The IPA lacks any specific symbol for the soundand more generally lacks any diacritic to indicate cross-sectional tongue shape (whetherflat or grooved) As a result a plethora of ad hoc transcriptions of slit-t can be found in theliterature Those known to us are the following

(2) [r] (Harris 1984 139 Lass 1987 267)

[s]

[]

[]

[s]

[si

[ri

mft]ft]

ft]

ft]

[ts]

[6]

(Harris 1990 285)

(Bertz 1987 45)

(Bertz 1987 45)

(Bertz 1987 45)

(Pilch 1959 420)

(Hogan 1934 18 Henry

(Conrick 198173)

(6Baoill 1990 161)1

(Wells 1982 429 Kallen

(IPA 1995 15)

(Hickey 1984a 234-235

(Henry 1958 123 Barry

(Henry 1958 123)

(Laver 1994 260)2

1958 123)

1994178)

1984b 675-676 1993 221-222 1995 123-124)

198168)

1 We have been informed by 6 Baoill (personal communication 25 ix 1995) that this perplexingtranscription mdash how can [t] be raised mdash is not his but was suggested by an editor of the volume in whichhis paper appears owing to the difficulty of setting his preferred transcription viz [ t] typographically

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

70 PANDELI ESKA BALL amp RAHILLY

Some of these transcriptions are simply idiosyncratic others are misleading or indeedinaccurate It can be seen from the above list that transcribers have generally used as basesymbols the following [r] [s] [t] and [8] We will now look at each of these in turn

[r]

This symbol represents a tap (or flap or flick in some terminology) which isessentially a very short sound Catford (1977129-30) states that the hit registered by aflick or by a passing transient flap has a measurable duration of the order of 1 to 3centiseconds and Ladefoged amp Maddieson (1996) cite Quilis (1981) who reports that themean closure duration of tokens of Spanish [r] which he measured was 20 ms As themean duration of Hiberno-English slit-t in the data presented above is 239 ms this clearlyrules out any consideration of the sound as a tap flap It would also be expected that analveolar tap would normally be formed with a complete closure on the alveolar ridge(though Laver 1994 reports the existence of tapped fricatives) Such a closure is notfound for slit-t [r] must therefore be discounted as the base symbol for this sound

[s]

The choice of [s] as a symbol is presumably motivated by an attempt to demonstratethe fricative nature of slit-t as well as its alveolar place of articulation We feel howeverthat this symbol is bound to be misleading as the IPA does not have any recogniseddiacritics to indicate the cross-sectional shape of the tongue This is a problem as the datapresented above suggest that the tongue shape for slit-t is flat (as for []) whereas for [s] itis grooved

[t]

With [t] the problem of cross-sectional tongue shape is less important The basesymbol [t] together with the lowering diacritic (ie [t ] = []) suggests a fricative mannerof articulation but not the grooved tongue shape of [s]

However the symbol [t] is ambiguous in terms of place of articulation On the onehand the Principles of the IPA (IPA 1949 17) state When it is desired to show inwriting that t d n are alveolar and not dental the retraction sign in the form - may beplaced underneath thus t d n This convention is still being adhered to by a number ofresearchers in the transcription of languages which contrast dental and alveolarconsonants For example Spajic et al (1996) use the retraction diacritic to transcribealveolar consonants in Toda (contrasting with a series of dental consonants) This sameexample is quoted in Ladefoged and Maddieson (1996 22) where the voiceless alveolarplosive of Toda is transcribed as [t]

2 Ladefoged and Maddieson (1996 144) also employ [8] to transcribe a voiceless flat alveolar fricative butdo not cite examples from southern Hiberno-English

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

68 PANDELI ESKA BALL amp RAHILLY

frame of one typical token of the fricative in the word met is shown in Figure 1

Analysis of the data revealed that contact was concentrated at the sides of the palateacross an area stretching from the front to the middle or back of the alveolar ridge A ratherbroad fricative channel was evident whose mean width was estimated to be 104 mmThis width was very similar to that estimated for tokens of [J] produced by the samespeaker (102 mm) but significantly greater than that for tokens of [s] (60 mm) TypicalEPG contact patterns for [s] and [J] are also shown in Figure 1 Finally the anterior edgeof the area of contact between the tongue and the palate was typically rather flat in shapefor slit-t with no contact at the sides of the palate in front of the major constrictionPandeli (1993) argues that this type of contact pattern is indicative of an apical articulation

In summary then the articulatory data reported in Pandeli (1993) suggest that for thisspeaker at least slit-t is an apical alveolar fricative formed with a broad central channelThe tongue shape which produces this type of fricative channel has been described in theliterature as eg flat (Pike 1943) and slit (Hockett 1955 Ladefoged 1971)

i IliUllMIIIlitilliiiitttiiilllll0231 Tine (sec) 0740

Figure 2 Spectrogram of met

In order to gain an impression of the durational characteristics of slit-t spectrogramswere made from acoustic recordings made at the same time as the EPG recordings Anexample spectrogram of the word met is shown in Figure 2 The duration of each of the 36tokens of slit-t was measured on the corresponding spectrogram and mean durationscalculated of the fricative in each of the four test words and in all words taken togetherFor comparison measurements were also made of the duration of [s] in 9 tokens each ofthe words lease mess mass and puss and of [J] in 9 tokens of leash mesh mash andpush These words were produced by the same speaker and recorded during the samesession as the slit-t words Mean durations of the three fricatives are shown in Table 1 Itcan be seen that the duration of all three fricatives is rather long reflecting their word-finalutterance-final position as well as a rather slow speaking rate Slit-t has a somewhatshorter mean duration (239 ms) than [s] (290 ms) and [f] (259 ms)

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 69

Table 1 Mean duration (ms) of Hiberno-English fricatives (9 tokens of each word)

slit-t

S

J

Neat

246

Lease

296

Leash

276

Met

245

Mess

284

Mesh

261

Mat

221

Mass

272

Mash

238

Put

243

306

Push

261

All

239

AH

290

All

259

4 Symbolising Slit-t

Given the physical characteristics of slit-t outlined above we will now consider howthe sound might best be transcribed The IPA lacks any specific symbol for the soundand more generally lacks any diacritic to indicate cross-sectional tongue shape (whetherflat or grooved) As a result a plethora of ad hoc transcriptions of slit-t can be found in theliterature Those known to us are the following

(2) [r] (Harris 1984 139 Lass 1987 267)

[s]

[]

[]

[s]

[si

[ri

mft]ft]

ft]

ft]

[ts]

[6]

(Harris 1990 285)

(Bertz 1987 45)

(Bertz 1987 45)

(Bertz 1987 45)

(Pilch 1959 420)

(Hogan 1934 18 Henry

(Conrick 198173)

(6Baoill 1990 161)1

(Wells 1982 429 Kallen

(IPA 1995 15)

(Hickey 1984a 234-235

(Henry 1958 123 Barry

(Henry 1958 123)

(Laver 1994 260)2

1958 123)

1994178)

1984b 675-676 1993 221-222 1995 123-124)

198168)

1 We have been informed by 6 Baoill (personal communication 25 ix 1995) that this perplexingtranscription mdash how can [t] be raised mdash is not his but was suggested by an editor of the volume in whichhis paper appears owing to the difficulty of setting his preferred transcription viz [ t] typographically

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

70 PANDELI ESKA BALL amp RAHILLY

Some of these transcriptions are simply idiosyncratic others are misleading or indeedinaccurate It can be seen from the above list that transcribers have generally used as basesymbols the following [r] [s] [t] and [8] We will now look at each of these in turn

[r]

This symbol represents a tap (or flap or flick in some terminology) which isessentially a very short sound Catford (1977129-30) states that the hit registered by aflick or by a passing transient flap has a measurable duration of the order of 1 to 3centiseconds and Ladefoged amp Maddieson (1996) cite Quilis (1981) who reports that themean closure duration of tokens of Spanish [r] which he measured was 20 ms As themean duration of Hiberno-English slit-t in the data presented above is 239 ms this clearlyrules out any consideration of the sound as a tap flap It would also be expected that analveolar tap would normally be formed with a complete closure on the alveolar ridge(though Laver 1994 reports the existence of tapped fricatives) Such a closure is notfound for slit-t [r] must therefore be discounted as the base symbol for this sound

[s]

The choice of [s] as a symbol is presumably motivated by an attempt to demonstratethe fricative nature of slit-t as well as its alveolar place of articulation We feel howeverthat this symbol is bound to be misleading as the IPA does not have any recogniseddiacritics to indicate the cross-sectional shape of the tongue This is a problem as the datapresented above suggest that the tongue shape for slit-t is flat (as for []) whereas for [s] itis grooved

[t]

With [t] the problem of cross-sectional tongue shape is less important The basesymbol [t] together with the lowering diacritic (ie [t ] = []) suggests a fricative mannerof articulation but not the grooved tongue shape of [s]

However the symbol [t] is ambiguous in terms of place of articulation On the onehand the Principles of the IPA (IPA 1949 17) state When it is desired to show inwriting that t d n are alveolar and not dental the retraction sign in the form - may beplaced underneath thus t d n This convention is still being adhered to by a number ofresearchers in the transcription of languages which contrast dental and alveolarconsonants For example Spajic et al (1996) use the retraction diacritic to transcribealveolar consonants in Toda (contrasting with a series of dental consonants) This sameexample is quoted in Ladefoged and Maddieson (1996 22) where the voiceless alveolarplosive of Toda is transcribed as [t]

2 Ladefoged and Maddieson (1996 144) also employ [8] to transcribe a voiceless flat alveolar fricative butdo not cite examples from southern Hiberno-English

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 69

Table 1 Mean duration (ms) of Hiberno-English fricatives (9 tokens of each word)

slit-t

S

J

Neat

246

Lease

296

Leash

276

Met

245

Mess

284

Mesh

261

Mat

221

Mass

272

Mash

238

Put

243

306

Push

261

All

239

AH

290

All

259

4 Symbolising Slit-t

Given the physical characteristics of slit-t outlined above we will now consider howthe sound might best be transcribed The IPA lacks any specific symbol for the soundand more generally lacks any diacritic to indicate cross-sectional tongue shape (whetherflat or grooved) As a result a plethora of ad hoc transcriptions of slit-t can be found in theliterature Those known to us are the following

(2) [r] (Harris 1984 139 Lass 1987 267)

[s]

[]

[]

[s]

[si

[ri

mft]ft]

ft]

ft]

[ts]

[6]

(Harris 1990 285)

(Bertz 1987 45)

(Bertz 1987 45)

(Bertz 1987 45)

(Pilch 1959 420)

(Hogan 1934 18 Henry

(Conrick 198173)

(6Baoill 1990 161)1

(Wells 1982 429 Kallen

(IPA 1995 15)

(Hickey 1984a 234-235

(Henry 1958 123 Barry

(Henry 1958 123)

(Laver 1994 260)2

1958 123)

1994178)

1984b 675-676 1993 221-222 1995 123-124)

198168)

1 We have been informed by 6 Baoill (personal communication 25 ix 1995) that this perplexingtranscription mdash how can [t] be raised mdash is not his but was suggested by an editor of the volume in whichhis paper appears owing to the difficulty of setting his preferred transcription viz [ t] typographically

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

70 PANDELI ESKA BALL amp RAHILLY

Some of these transcriptions are simply idiosyncratic others are misleading or indeedinaccurate It can be seen from the above list that transcribers have generally used as basesymbols the following [r] [s] [t] and [8] We will now look at each of these in turn

[r]

This symbol represents a tap (or flap or flick in some terminology) which isessentially a very short sound Catford (1977129-30) states that the hit registered by aflick or by a passing transient flap has a measurable duration of the order of 1 to 3centiseconds and Ladefoged amp Maddieson (1996) cite Quilis (1981) who reports that themean closure duration of tokens of Spanish [r] which he measured was 20 ms As themean duration of Hiberno-English slit-t in the data presented above is 239 ms this clearlyrules out any consideration of the sound as a tap flap It would also be expected that analveolar tap would normally be formed with a complete closure on the alveolar ridge(though Laver 1994 reports the existence of tapped fricatives) Such a closure is notfound for slit-t [r] must therefore be discounted as the base symbol for this sound

[s]

The choice of [s] as a symbol is presumably motivated by an attempt to demonstratethe fricative nature of slit-t as well as its alveolar place of articulation We feel howeverthat this symbol is bound to be misleading as the IPA does not have any recogniseddiacritics to indicate the cross-sectional shape of the tongue This is a problem as the datapresented above suggest that the tongue shape for slit-t is flat (as for []) whereas for [s] itis grooved

[t]

With [t] the problem of cross-sectional tongue shape is less important The basesymbol [t] together with the lowering diacritic (ie [t ] = []) suggests a fricative mannerof articulation but not the grooved tongue shape of [s]

However the symbol [t] is ambiguous in terms of place of articulation On the onehand the Principles of the IPA (IPA 1949 17) state When it is desired to show inwriting that t d n are alveolar and not dental the retraction sign in the form - may beplaced underneath thus t d n This convention is still being adhered to by a number ofresearchers in the transcription of languages which contrast dental and alveolarconsonants For example Spajic et al (1996) use the retraction diacritic to transcribealveolar consonants in Toda (contrasting with a series of dental consonants) This sameexample is quoted in Ladefoged and Maddieson (1996 22) where the voiceless alveolarplosive of Toda is transcribed as [t]

2 Ladefoged and Maddieson (1996 144) also employ [8] to transcribe a voiceless flat alveolar fricative butdo not cite examples from southern Hiberno-English

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

70 PANDELI ESKA BALL amp RAHILLY

Some of these transcriptions are simply idiosyncratic others are misleading or indeedinaccurate It can be seen from the above list that transcribers have generally used as basesymbols the following [r] [s] [t] and [8] We will now look at each of these in turn

[r]

This symbol represents a tap (or flap or flick in some terminology) which isessentially a very short sound Catford (1977129-30) states that the hit registered by aflick or by a passing transient flap has a measurable duration of the order of 1 to 3centiseconds and Ladefoged amp Maddieson (1996) cite Quilis (1981) who reports that themean closure duration of tokens of Spanish [r] which he measured was 20 ms As themean duration of Hiberno-English slit-t in the data presented above is 239 ms this clearlyrules out any consideration of the sound as a tap flap It would also be expected that analveolar tap would normally be formed with a complete closure on the alveolar ridge(though Laver 1994 reports the existence of tapped fricatives) Such a closure is notfound for slit-t [r] must therefore be discounted as the base symbol for this sound

[s]

The choice of [s] as a symbol is presumably motivated by an attempt to demonstratethe fricative nature of slit-t as well as its alveolar place of articulation We feel howeverthat this symbol is bound to be misleading as the IPA does not have any recogniseddiacritics to indicate the cross-sectional shape of the tongue This is a problem as the datapresented above suggest that the tongue shape for slit-t is flat (as for []) whereas for [s] itis grooved

[t]

With [t] the problem of cross-sectional tongue shape is less important The basesymbol [t] together with the lowering diacritic (ie [t ] = []) suggests a fricative mannerof articulation but not the grooved tongue shape of [s]

However the symbol [t] is ambiguous in terms of place of articulation On the onehand the Principles of the IPA (IPA 1949 17) state When it is desired to show inwriting that t d n are alveolar and not dental the retraction sign in the form - may beplaced underneath thus t d n This convention is still being adhered to by a number ofresearchers in the transcription of languages which contrast dental and alveolarconsonants For example Spajic et al (1996) use the retraction diacritic to transcribealveolar consonants in Toda (contrasting with a series of dental consonants) This sameexample is quoted in Ladefoged and Maddieson (1996 22) where the voiceless alveolarplosive of Toda is transcribed as [t]

2 Ladefoged and Maddieson (1996 144) also employ [8] to transcribe a voiceless flat alveolar fricative butdo not cite examples from southern Hiberno-English

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 71

On the other hand the following statement is made in the Report on the 1989 KielConvention of the IPA (IPA 198970)

The dental and postalveolar columns will not be divided by a gridline fromthe alveolar one except where distinct symbols are approved The dental diacriticmay be used to distinguish dental sounds and post-alveolar ones may bedistinguished by use of the diacritic for retraction

This principle is reiterated in the Preview of the IPA Handbook (IPA 19957)

Note that except in the case of fricatives only one symbol is provided fordental alveolar postalveolar if necessary these three places can bedistinguished by the use of extra marks or diacritics to form compositesymbols For example the dental alveolar postalveolar nasals can berepresented as [n n n] respectively

An obvious ambiguity exists here The retraction diacritic may be used to indicate analveolar as opposed to a dental consonant or a postalveolar as opposed to an alveolarconsonant The plain symbol [t] may correspondingly be taken to represent either a dentalor an alveolar consonant [t] (= []) is therefore ambiguous in terms of place ofarticulation As we argue below this suggests the need for a separate alveolar diacritic

[6]

This symbol used with the retraction diacritic ([6]) probably constitutes the mostsatisfactory transcription of slit-t which has so far been suggested Fricative manner ofarticulation is inherent in the base symbol and a flat cross-sectional tongue shape isimplied The retraction diacritic also gives an indication of the place of articulation of theconsonant telling us that the sound is articulated at some point behind the interdentalposition The problem though is that we are left to guess where exactly this might beRetraction could be to a point just behind the teeth (so that the sound is no longerwiferdental) or it could be to a point as far back as the alveolar ridge

Clearly what is required for a less ambiguous transcription of slit-t is a diacritic thatindicates alveolar and only alveolar place of articulation As we noted earlier althoughthe IPA has a diacritic to indicate dental place and a method (albeit ambiguous) ofindicating postalveolar place it does not have an alveolar diacritic We have shown thatthis can lead to problems in the transcription of normal speech and researchers workingon the transcription of disordered speech have encountered similar problems (see eg Ballet al 1996) For this reason the extlPA for disordered speech (Duckworth et al 1990)has adopted the subscript equals sign as the diacritic to denote alveolar place bothsegmentally and prosodically (ie as a supralaryngeal voice quality)

In our view there is no reason why symbols of the extlPA should not be employed forthe transcription of the sounds of normal speech Under current IPA practice [t]mdashwhichmay be preferred by phonologists for the sake of phonemic transparency (in SouthernHiberno-English) and [6]mdashwhich is preferable from a general phonetic point of view areas close as it is possible to come to transcribing slit-t adequately Neither of these

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

72 PANDELI ESKA BALL amp RAHILLY

transcriptions however provides an unambiguous representation of the articulation of thesound A more precise transcription can be made if the extlPA alveolar diacritic is usedWe therefore suggest in the interests of transcriptional precision that the SouthernHiberno-English slit-t fricative is transcribed as [9]

5 Variation in alveolar flat fricatives

Our recommendation that southern Hiberno-English slit-t be transcribed as [0] is basedon detailed consideration of data from only one speaker It seems probable however thata certain amount of variation exists between Hiberno-English speakers in the way in whichslit-t is realised phonetically The symbol we recommend may therefore not always be themost appropriate Jeff Kallen (personal communication 27 ii 1996) suggests for instancethat some speakers produce the sound as an affricate in which case [t0] might be a moresuitable transcription The inconsistency noted in Section 3 in the way in which slit-t hasbeen transcribed may partly reflect this between-speaker variation in the production of thesound It may however also reflect the impressionistic nature of the transcriptions whichwere made without the aid of instrumental data

Having dealt with the southern Hiberno-English slit-t in some detail we would likefinally to turn our attention to alveolar flat fricatives in another language mdash Icelandic mdashand consider how these might best be transcribed Icelandic has both fortis and lenisalveolar flat fricatives that are normally represented by the symbols [0 8] in broadtranscription The narrower transcription [8 5] is suggested by Laver (1994) toemphasize the fact that the fricatives are alveolar rather than dental or interdental(Petursson 1971) Laver therefore uses the symbol [0] to represent both the Icelandic fortisalveolar flat fricative and southern Hiberno-English slit-t The two sounds do not appearto us to be auditorily equivalent however and there is also evidence to suggest acorresponding articulatory difference between them

Petursson presents data derived from both cineradiography and static directpalatography for the Icelandic flat fricatives which he transcribes as [fgt 9] and thegrooved fricative [s] He does not provide channel width measurements from hispalatograms but if these are compared with the palatogram of Hiberno-English slit-tshown in Figure 1 it can be seen that for the latter there is more peripheral contact and amore obvious narrowing of contact in the alveolar region This suggests that for theIcelandic flat fricatives the tongue is lower in the mouth than for Hiberno-English slit-tPetursson actually provides a measure of the distance between the tongue surface and theroof of the mouth for the Icelandic fricatives and reports it to be 85-11 mm for [fgt 5]This is considerably more than the distance reported for [s] (4-6 mm) and this againsuggests that the tongue is rather low in the mouth for the Icelandic flat fricatives In thedescription of Icelandic given in the Principles of the IP A the lenis member of the pair is infact described as a frictionless continuant (IPA 1949 28)

The evidence suggests then that the Icelandic flat alveolar fortis fricative is differentfrom Hiberno-English slit-t and we would argue therefore that the two sounds should notbe represented in the same way when making a narrow phonetic transcription The

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 73

difference between them could be captured either by using a lowering diacritic with thesymbol for the Icelandic flat fricative or a raising diacritic for slit-t It is alreadysanctioned practice within the IPA (IPA 1995) to use a lowering diacritic to markapproximant realisations of fricatives and as we have seen Icelandic [5] has beendescribed as a frictionless continuant or approximant We therefore suggest thetranscription [6J for the Icelandic fricative and [8] for slit-t

6 Conclusion

No system of phonetic notation can ever be expected to have symbols and diacritics todenote every possible aspect of articulation It might appear superficially surprising that inthe recent adoption by the IPA of diacritics to distinguish tip and blade articulations (IPA1989) the Association did not also approve symbolisations to denote tongue body shape aswell However any increase in the alphabet has of course to be justified by a relativelywidespread use of the sounds in question as well as a need to contrast them in transcriptionfrom other similar sounds As this account has shown slit alveolar fricatives are seldomrecorded

Nevertheless in a narrow transcription of southern Hiberno-English we may well needto distinguish between slit alveolar and dental fricatives as well as grooved alveolarfricatives In the absence of a slit-grooved diacritic we have demonstrated that the use ofan alveolar diacritic from the extlPA symbol system is a good solution as it is not onlyarticulatorily accurate but avoids the ambiguities of other proposals

Acknowledgements

We should like to thank Jeff Kallen for much discussion and assistance in locating examples ofvarious transcriptional practices It should not be inferred that he necessarily shares our opinions

References

BALL MJ RAHILLY J amp TENCH P (1996) The Phonetic Transcription of DisorderedSpeech San Diego Singular Publishing

BARRY MV (1981) The southern boundaries of northern Hiberno-English speech InMV Barry (ed) Aspects of English Dialects in Ireland vol 1 52-93 BelfastQueens University of Belfast

BERTZ S (1987) Variation in Dublin English Teanga 7 35-53CATFORD J (1977) Fundamental Problems in Phonetics Edinburgh Edinburgh

University PressCONRICK M (1981) Error analysis of Irish students learning French Teanga 2 70-75DUCKWORTH M ALLEN G HARDCASTLE W and BALL MJ (1990) Extensions to

the International Phonetic Alphabet for the transcription of atypical speech ClinicalLinguistics and Phonetics 4 273-80

HARRIS J (1984) English in the north of Ireland In P Trudgill (ed) Language in theBritish Isles Cambridge Cambridge University Press

HARRIS J (1990) Segmental complexity and phonological government Phonology 7 255-300

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

74 PANDELI ESKA BALL amp RAHILLY

HENRY PL (1958) A linguistic survey of Ireland Preliminary report Lochlann 1 49-208

HlCKEY R (1984a) Coronal segments in Irish English Journal of Linguistics 20 233-250

HlCKEY R (1984b) Phonotactically conditioned alternation Instances from Old HighGerman and Irish English Linguistics 22 673-686

HlCKEY R (1986) Possible phonological parallels between Irish and Irish EnglishEnglish World-Wide 7 1-21

HlCKEY R (1993) The beginnings of Irish English Folia Linguistica Historica 14 213-238

HlCKEY R (1995) An assessment of language contact in the development of IrishEnglish In J Fisiak (ed) Linguistic Change under Contact Conditions 109-130Berlin Mouton de Gruyter

HlCKEY R (1996) Identifying dialect speakers the case of Irish English In H Kniffa(ed) Recent Developments in Forensic Linguistics Frankfurt Peter Lang

HOCKETT C (1955) A Manual of Phonology Baltimore Waverley PressHOGAN JJ (1934) An Outline of English Phonology chiefly for Irish Students Dublin

amp Cork The Educational Company of Ireland LtdIPA (1949) The Principles of the International Phonetic Association London IPAIPA (1989) Report on the 1989 Kiel Convention Journal of the International Phonetic

Association 19 67-80IPA (1995) Preview of the IPA Handbook Journal of the International Phonetic

Association 25 3-47KALLEN JL (1994) English in Ireland In R Burchfield (ed) The Cambridge History

of the English Language vol v English in Britain and Overseas Origins andDevelopment Cambridge Cambridge University Press

LADEFOGED P (1971) Preliminaries to Linguistic Phonetics Chicago The University ofChicago Press

LADEFOGED P (1990) What do we symbolize Thoughts prompted by bilabial andlabiodental fricatives Journal of the International Phonetic Association 20 32-36

LADEFOGED P amp MADDIESON I (1996) The sounds of the worlds languages OxfordBlackwell

LASS R (1987) The Shape of English Structure and History London JM Dent andSons Ltd

LAVER J (1994) Principles of Phonetics Cambridge Cambridge University PressNl GHALLCH6IR C (1981) Aspects of bilingualism in NW Donegal In MV Barry

(ed) Aspects of English Dialects in Ireland 142-170 Belfast Queens Universityof Belfast

OBAOILL DP (1990) Language contact in Ireland The Irish phonological substratumin Irish-English In JA Edmondson C Feagin amp P Miihlhausler (eds)Development and Diversity Language Variation across Time and Space Festschriftfor Charles-James N Bailey 147-172 Arlington TX Summer Institute ofLinguistics University of Texas at Arlington

PANDELI H (1993) The Articulation of Lingual Consonants An EPG Study PhDthesis University of Cambridge

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available

HIBERNO-ENGLISH SLIT-T 75

PETURSSON M (1971) Etude de la realisation des consonnes islandaises fgt 5 s dans laprononciation dun sujet islandais a partir de la radiocinematographie Phonetica 2 3 203-216

PIKE KL (1943) Phonetics Ann Arbor The University of Michigan PressPILCH H (1959) Neue Wege der englischen Phonetik Anglia 77 407-428SPAJIC S LADEFOGED P amp BHASKARARAO P (1996) The trills of Toda Journal of

the International Phonetic Association 26 1-21STONE M FABER A RAPHAEL LJ amp SHAWKER TH (1992) Cross-sectional

tongue shape and linguopalatal contact patterns in [s] [J] and [1] Journal ofPhonetics 20 253-270

TILLING P (1981) Age-group variation in the speech of Kinlough Co Leitrim In MVBarry (ed) Aspects of English Dialects in Ireland 96-105 Belfast QueensUniversity of Belfast

WELLS JC (1982) Accents of English vol ii The British Isles CambridgeCambridge University Press

at httpwwwcambridgeorgcoreterms httpdxdoiorg101017S0025100300005430Downloaded from httpwwwcambridgeorgcore University Libraries - Virginia Tech on 13 Oct 2016 at 145608 subject to the Cambridge Core terms of use available