Phonemic Inventory and Transcription. The symbols in Table 1 stand for the distinctive speech sounds (phonemes) of the Nuristâni languages.
|tip up||tip up or down||tip down||tip up||tip down|
|apico- or lamino-
|Affricates||ć or ç||č or c||c̣ or C|
||ǰ or j||j̣ or J||[vcd]|
|Stops||p||t||ṭ or T||k||(q)||(ˀ or >)|
|b||d||ḍ or D||g||[vcd]|
|Spirants||(f)||s||š or $||ṣ or S||(x)||(ḥ or H)||(h)|
|v||z||ž or 2||ẓ or Z||(ǧ or G)||(ˁ or <)||[vcd]|
|forward||ṛ, R, or R|
|nasal||ň or ñ||[nas]|
|oral||ř or R|
|Nasal Stops||m||n||ṇ or N||ŋ or Ñ||[nas]|
` ^ ´
|emphatic full-drop||? or !|
Where a phoneme is represented by multiple symbols in Table 1, the first symbol listed is the currently preferred one. Such symbols are based on the 16-bit Unicode system for encoding characters, which allows depiction of symbols used in traditional linguistic transcription. Older symbols, based on the restricted character sets of 7- and 8-bit encoding, are listed to the right of the preferred symbol. They are being phased out on this website.
Symbols enclosed in parentheses represent foreign sounds that are not fully integrated into the traditional sound systems.
The symbols [vcd] (voiced), [nas] (nasalized), and [frn] (fronted) represent coproduced components of the sounds listed to their left in the row.
Table 2 presents symbols for differing degrees of transition between or within lexical items in the speech stream. Additionally, in grammatical examples the hyphen (-) marks morpheme boundaries within a word.
|Tighter ← Degree of Transition → Looser|
Table 3 shows the indication of expressive lengthening by italic font style.
|Consonantal||Vocalic Rising||Vocalic Two-Topped|
Abbreviations. Abbreviations of language and dialect names used below include:
|Vâsʹi vari||V.||V.seć = dialect of sʹeć; V.ṣupu = dialect of ṣupʹu; V.uṣüt = dialect of uṣüt|
|Kâmkʹata vari||K.||K.kt = kâtʹa vari; K.ktw = Western kâtʹa vari; K.kte = Eastern kâtʹa vari; K.ktv = kâtʹa vari of ktʹivi; K.km = kâmvʹiri|
|Âṣkuňu-Saňu vi:ri||A.||A.s = saňu dialect; A.m = mâse:gal dialect|
|Kalaṣa-alâ||Kal.||Kal.n = dialect of nišeigrâm; Kal.v = dialect of vägal; Kal.a = dialect of amešdeš|
|Tregâmi||Tr.||Tr.g = dialect of gambir|
Sources and Reliability. Data from K.km, K.ktv, Kal.n, Kal.v, A.s, and V.ṣupu were collected in the field by this author (Strand). Data from other dialects are from Morgentierne (1929a, 1934a, 1949, 1952, 1954, and unpublished), Grjunberg (1980) for K.ktw, and Buddruss (2005) for V.uṣüt, unless otherwise noted. Phonemic analysis of raw field data is complete for K.km. and mostly complete for K.ktv, K.ktw, Kal.n, A.s, and V., but a few issues remain unresolved in each of these dialects. Morgenstierne's data are largely pre-phonemic and sometimes leave unresolved the phonemic status of his recorded vowel phones, such as [ə] vs. [a], [ā] vs. [ɔ] vs. [o], and vocalic length.
When cited on this website, forms from other researchers are respelled to fit the system given here. Notable are the substitution here of a for Buddruss's and Morgenstierne's "ə" and Grjunberg's "i", â for "a", and i for Grjunberg's "ī". Also, ć and ź are used in place of "c" and "j" for the dental affricates.
Principles of Transcription. The system of transcription used here is phonemic, in which each symbol represents a unit of speech that contrasts with all other units in the system. It is important to realize that actual speech sounds are not inherent in a phonemic symbol; rather, each symbol represents a contrastive unit of speech, which may have differing pronunciations from language to language or in different positions within utterances of a single language. To indicate actual speech sounds (phones), conventional phonetic symbols are used, enclosed in brackets according to usual linguistic practice (see below). For example, note the different pronunciations of the phonemic symbol a given below.
Underlying each phoneme are articulatory processes (components) that produce a target sound. Such processes are indicated by abbreviations for the process name, enclosed in brackets; e.g., [frn] (fronted) for pushing the blade of the tongue forward to the front of the mouth, [nas] (nasalized) for lowering the velum to open the nasal passage, and [vcd] (voiced) for adducting the vocal folds to produce the vibration known as "voicing."
One goal of phonemic transcription is to minimize redundancy in the transcriptional system. This means that if a component of one speech sound always occurs in conjunction with a component of another, usually adjacent sound, that component's indication in the transcription may be eliminated because its occurrence is predictable from the occurrence of the determining sound.
For example, the obstruent consonants (see Table 1) may be contrastively voiced or voiceless at the beginning of a word, as in K.km pʹut 'path' vs. bʹut 'cooked rice', tʹü 'thou' vs. dʹü 'two', ṣʹia 'scrape into!' vs. ẓʹia 'black', c̣ʹuc̣ 'mother-in-law' vs. j̣ʹuc̣ 'name of a pre-Islamic god', etc. The difference in each pair is marked by the process of voicing ([vcd]) in the voiced consonant, while the voiceless consonant is unmarked for voicing. In K. sequences of consonants that contain a series of obstruents or obstruents plus nasal stop are either all voiced or all voiceless. The voicing of the entire sequence is determined by the voicing of the final obstruent or nasal stop in the sequence, so that the voicing of the preceding consonants in the sequence may be indicated conventionally by a symbol for the unmarked (voiceless) consonant, with the understanding that the unmarked consonants automatically take on the voicing of the final one. This is seen in sequences of words or morphemes in which the second item begins with a voiced consonant. The final consonant of the first item is written with the symbol for a voiceless (unmarked) consonant, which is automatically pronounced with the anticipated voicing of the second consonant; e.g., in K.km iʹuṣ 'demon' plus dâmʹu 'wind' produces the compound word iâṣdʹamu 'whirlwind', in which the ṣ is pronounced as ẓ; similarly, vʹoṭ 'rock' plus muřʹi 'clay' produces the compound word vâṭmʹuři 'rock and clay mixture', in which the ṭ is pronounced as ḍ. To write these as xiâẓdʹamu and xvâḍmʹuři would redundantly indicate the automatically anticipated voicing of the second consonant in the first consonant of the sequence.
Another goal of phonemic transcription is to use a single symbol for each phoneme, while minimizing the total number of symbols in the system. Thus we use u to indicate the sound written with two letters in English ("oo") or French ("ou"), and spellings such as "Nooristan" or "Nouristan" are to be avoided as archaically unphonemic. However, this goal is overridden by convention in certain cases; e.g., nasalized vowels are written as vowel plus ˜ rather than with the additional symbols ã, õ, etc., but the nasal stops are written with m, n, etc., rather than as p˜, t˜, etc.
Comparison of Nuristâni Sounds with English and IPA Standards. To assist the reader in pronouncing Nuristâni sounds, approximate equivalents in Midwestern American English are given in the following descriptions. Symbols for individual phones appear in brackets and broadly follow the conventions of the International Phonetic Association (IPA), as modified by traditional practices of American linguists.
Default Speech Posture and Processes. Each dialect has a default posture of the vocal tract that is assumed during speech. In all dialects the front of the larynx is tensed, producing anterior voicing in the glottis. The tension spreads from the larynx up through the root of the tongue, pushing the tongue closer to the roof of the mouth and farther forward. The process that produces this coordinated raising of the larynx and the tongue is called laryngeal-lingual tensing.
Default Degree of Laryngeal-Lingual Tensing. The default degree of laryngeal-lingual tensing tension in each dialect differentiates it from others. Default laryngeal-lingual tension is lowest in Kal.n, somewhat higher in Kal.v and A., and highest in K. and V. Increased tension caused ancestral vowels to become progressively more close in Kal.v and A., reaching maximal closeness in K. and V. This development can be seen in the evolution of the ancient unmarked vowel a into close [ɨ] in K., V., A., and Kal.v, as described below. The distribution of the close pronunciation of a may be interpreted as strengthened tenseness that was innovated in the Northern Nuristâni languages K. and V. and spread to the neighboring dialects A. and Kal.v.
Strengthened tenseness in the Northern Nuristâni languages also underlies the extensive use of tensed expressive lengthening in K.km, while in Kal.n expressive lengthening is less ubiquitous.
|kârṭʹa 'fat'||biliʹuk kârṭa 'very fat'||öčü karaṭa 'very fat'|
|kârṭʹa ([ṭ:]) 'extremely fat'||biliʹuk kârṭa ([l:]) 'really fat!'||öčü karaṭa ([t:š]) 'really fat!'|
Fronting and Prognathizing. In Kal. and, to a much lesser extent, K., fronting is anticipated from the vowel of a word-final syllable to precdeding vowels in the word, as in:
|ḍabala 'boy'||soi (Kal.v) 'sun'||ʹutra 'sonless male'|
|ḍäbäli 'girl'||sö (Kal.n) 'sun'||ʹütri 'sonless female'|
"Prognathizing" is a jutting-out of the jaw, pushed by the tip of a fronted tongue against the lower teeth. It is the default posture in A. and Kal. and is also seen in the pronunciation of other languages of the region, notably Pashto.
Accent. Certain vowels in the speech stream are accented by stronger tensing of the anterior vocal folds, producing a raised pitch above that of unaccented vowels. At the end of a declarative utterance, the tensing drops off abruptly, causing a drop in pitch and voicing. The default intonational pattern for an utterance is therefore 1 2 0 , indicating normal followed by accented followed by dropped pitch. Because of strengthened tenseness in the Northern Nuristâni languages, the actual range between accented and dropped pitch is greater in those languages than in A.s and Kal.n.
Voicing. Voicing is produced according to the default posture described above: the anterior portion of the vocal folds ("vocal cords") are tensed to the point where they can vibrate in the passing airstream. The tensing raises the larynx and extends throughout its front upward to the root of the tongue. At the end of declarative utterances the tensing drops off, and the larynx returns to a relaxed position, which precludes final voicing in K.km. Voicing is never produced with the posterior portion of the vocal folds, as it is in the neighboring Indo-Aryan languages or in the Germanic languages.
In all dialects voicing is anticipated from a final consonant in a sequence of consonants to the preceding consonants of the sequence, except that nasal stops are always voiced; thus:
|o:sṭ ([ṣṭ]) 'eight'||osṭ ([ṣṭ]) 'eight'||ʹusṭ ([ṣṭ]) 'eight'||es ([ṣṭ]) 'our'|
|o:sṭviši ([ẓḍv]) 'eight score'||osṭviši ([ẓḍv]) 'eight score'||ʹuṣvići ([ẓv]) 'eight score'||es zʹâṭ ([zz]) 'our kinsman'|
Vowels. In at least the Northern Nuristâni languages vowels are fully articulated before they are voiced. Common to all Nuristâni languages are the phonemic contrasts between the vowels a, â, u, o, i, and e. To these are added the fronted vowels ü in Kal., K.km, K.kte, and V., ö in Kal. and V., and ä in Kal. and Tr. Marginally phonemic is a backed, lip-rounded vowel å found in Tr. and K.km.
a. a represents the basic vowel in each language, etymologically descendant from ancient Indo-Irânian *a. In A., K., and V. it is phonetically a close central vowel [ɨ] (like "i" in English "it", but further back), while in Kal. and Tr. (probably) it ranges from a half-opened central vowel ([IPA ([ɐ], like "u" in English "cut" or the vowel [ə] of English unstressed "the") to close [ɨ] in word-final position in Kal.v. In K. it is fronted to [i] (like English "ee") after lamino-alveolar consonants; thus cʹa 'how many' is pronounced [či]. In K.km an a after a vowel lengthens the vowel. Examples are:
|sa ([sˈɐ]) 'that'||sa ([sˈə]) 'that'||ga ([gˈɨ]) 'he went'||sʹa ([sˈɨ]) 'year'||cʹa ([čˈi]) 'how many'||dʹâa ([dˈaː]) 'wood'||tʹa ([tˈɨ]) 'river'|
|mać ([mˈɐts]) 'fish'||sas ([sˈɨs]) 'you are'||drʹaṣ ([drˈɨṣ]) 'handspan'||kšʹa ([kšˈi]) 'do it!'||dʹâ ([dˈa]) 'take!'||kʹas ([kˈɨs]) 'who'|
â. â is an open central vowel ([a]), like "a" in English "father." It contrasts with a by being articulated with the jaw more open. Examples are:
|gâ 'cow'||gâ 'cow'||gâ 'cow'||dʹâ 'take!'||iʹâ 'father'|
|bâš 'twelve'||bâš 'twelve'||bâs 'steam'||gʹâṣ 'billy-goat'||zʹâṭ 'kinsman'|
o. o is a lip-rounded, mid-height, non-fronted vowel with degree of closure ranging from somewhat close ([ʊ], like English "oo" in "book") in A., to mid ([o], like English "o" in "show," but without increasingly tenser lip-rounding) in V., Kal. and Tr., to slightly open [o̞] in K. Examples are:
|šo 'sand'||to 'thee'||to 'thee'||gʹo 'cow'||ṣʹo˜ 'soul'|
|doš 'ten'||dos 'yesterday'||dos 'yesterday'||drʹos 'grape'||liʹok 'both'|
u. u is a lip-rounded, close, central to back vowel, somewhat fronted in A.([ʉ]), as opposed to the other languages, where it is pronounced [u], like English "oo" in "noon." The different pronounciations reflect different origins: A. u comes from Indo-Iranian u, while in K. u derives from earlier o and au. Examples are:
|ṣu 'six'||ṣu 'six'||du 'two'||dʹu 'door'||kʹu 'stool'|
|dus 'yesterday'||dus ([dọs]) 'yesterday'||puc̣ 'flea'||dʹus 'yesterday'||lʹust 'hand'|
ü. ü is a lip-rounded, close, fronted vowel ([ü] or IPA [y]), which occurs in V., K.km, K.kte, and Kal. Its pronunciation is close to French "u" or German "ü." In K.ktv it has lost its phonemic distinctiveness by dissimilating into the sequence iu. Examples are:
|dü 'two'||dʹü 'two'||lʹü 'two'|
|püć 'pine'||pʹüs 'lost'||lʹüšt 'daughter'|
i. i is a close front vowel ([i]) in all languages, similar to English "ee" in "see." Examples are:
|bi 'seed'||i 'this'||di 'go'||bʹi 'seed'||ǰʹi 'head'|
|kiš 'what'||źim 'snow'||drist 'handspan'||pʹiš 'flower'||zʹir 'heart'|
e. e is a mid-height fronted vowel ([e], similar to the vowel in English "take"), except that it is somewhat lowered and retracted in K.kt ([ɛ̠], similar to English "e" in "get"). Examples are:
|bre 'flour'||tre 'three'||vřei 'flour'||břʹe 'flour'||pʹe 'down into'|
|deš 'village'||žei 'mother'||čeč 'wicker platter'||ṣʹeć 'sixteen'||wʹârek 'house'|
ä. ä is an open, fronted vowel ([æ], as in English "bad.") that appears in Kal. and Tr. Examples are:
|dä 'beard'||sä 'sun'|
|kä˜st 'husband's younger brother'||myär 'cloud'|
ö. ö is a lip-rounded, mid-opened, fronted vowel ([ø], similar to German "ö" or French "oe") that occurs in Kal. and, to a lesser extent, V. Examples are:
|zö 'heart'||liö 'wood'|
|gröṣ 'billy-goat'||sülmʹöč (a pre-Islamic goddess)|
å. å is a lip-rounded, open back vowel ([ɒ] or [ɔ]), similar to the vowel in English "caught." It occurs in Tr. and K.km, but its phonemic status as distinct from â or o is unclear in Tr. In K.km. it is a marginally phonemic sound used to represent Fârsi (Dari) â (phonetically [ɒ]) in loanwords. Examples are:
|gadå 'donkey'||zåbʹit 'sergeant'|
|dåš 'ten'||nåqʹåm 'failed'|
Nasalized Vowels. Nasalized vowels are indicated by a following ˜ . They appear in all dialects except K.ktw, where a raised velum has closed off the nasal passage to leave only oral vowels. Nasalization derives from a reduction of ancient *n. Examples are:
|ṣö˜ 'soul'||žu˜ 'me'||kuː˜ 'where'||dʹo˜ 'debt'||pi·ʹa˜ 'son'|
|ṣe˜š 'sixteen'||po˜č 'five'||břoː˜ć 'meadow'||břʹu˜ć 'meadow'|
Lengthened Vowels. In some dialects vowel length is distinctive; it is represented by postvocalic ː in A., V. and Kal.v, and by postvocalic a in K.km. Compare the lengthened vowels of the first row with the simple vowels of the second row:
|Kalaṣa.vägal||Âṣkuňu.saňu||Kâmkata.kâmviri||Vâsi.uṣüt||iː 'this one [oblique case]'||bâːs 'twelve'||ćʹoa 'branch'||lʹâːst 'they had'|
|i 'this one'||bâs 'vapor'||ćʹo 'edible greens'||lʹâst 'they have'|
Position and Representation of Accent. Accented vowels are produced with an anterior tensing of the vocal folds and larynx, causing a raised pitch on the accented vowel. Accent is indicated by ʹ before a vowel in K. and V., the two languages in which accentual position is distinctive. Compare the position of accent in each row:
|ʹâćaň 'come!' [plural]||ʹu tia 'stand up!'||ʹimi 'wife's brother'|
|âćʹaň 'carry on your backs!' [plural]||utʹia 'wait!'||irʹi 'co-wife'|
|`âćaň 'come!' [plural]||`u tia 'stand up!'||`imi 'wife's brother'|
|^âćaň 'carry on your backs!' [plural]||^utia 'wait!'||^iri 'co-wife'|
|´âpdulo+ǰon 'Abdullâh Jân'||´âmir+âpdurâhmon+xon 'Âmir Abdur-Rahmân Khân'|
In the remaining languages (including Tr.?) accent falls automatically on a word's final syllable (excluding inflectional or enclitic suffixes) and is not written.
Non-Distinctive Vocalic On-Glides. In K. and V. close vowels are articulated before they are vocalized, so that initial u, ü , and i have homorganic on-glides, non-phonemic in K.km, but promoted to phonemic status in other dialects through later vowel changes (cf. K.km ʹušpa with initial [w] but K.ktv vašʹup 'horse'). In Vâsi it is not clear whether initial onglides contrast with initial v and y before close vowels. Examples include:
|uřʹa ([wu]) 'quail'||uřʹa ([wu]) 'quail'||ʹut ([wu]) 'pre-Islamic priest' (V.ṣupu)|
|uṣʹa ([wu]) 'medicine'||vaṣʹa ([βɨ]) 'medicine'||wuṣʹu ([wu]?) 'medicine' (V.seć)|
|ʹušpa ([wu]) 'horse'||vašʹup ([βɨ]) 'horse'|
In Kâmkata.kâmviri the lip-rounding of initial u is fully articulated before the onset of voicing, producing a labial on-glide ([w], IPA [u̯]) to the vowel. This on-glide is distinct from the labio-dental spirant v found in, e.g., vʹu 'daughter-in-law'. Similarly, the tongue-fronting of initial i or ü is fully articulated before the onset of voicing, producing a palatalized on-glide ([y], IPA [i̯]) to the vowel, as in ʹi (IPA [i̯i]) 'go!' or ʹü (IPA [i̯y]) 'louse'. This on-glide is distinct from the more tensely articulated prevocalic i (a lamino-alveolar approximant before vowels, IPA [j]) found in, e.g., iʹu 'eat!'. Before initial ü the palatal fronting precedes both lip-rounding and voicing, and the palatal on-glide prevails.
Consonants. Consonants are produced by varying degrees of stricture to the airstream flowing out beyond the vocal folds. Each degree of stricture adds a corresponding degree of acoustic noise to the consonantal sound, except that nasalization ([nas]) adds a harmonic to the voicing process. A low degree of stricture allows voicing to be maintained during the consonant; a high degree of stricture above the glottis causes voicelessness in obstruents by reducing the airflow through the glottis to a rate incapable of sustaining vocal-fold vibration. In K.km the degree of stricture drops after the onset, leaving single intervocalic consonants to absorb the voicing of surrounding vowels. At the end of an utterance consonants are fully released, while the larynx drops and slackens the vocal folds toward reduced voicing of final consonants in Kal.n and K.ktv and full voicelesness in K.km.
The degrees of airstream stricture determine contrasts between stops (oral and nasalized), affricates, spirants, flaps, laterals, and approximants (see Table 1). Contrasts of voiced vs. voiceless stops, affricates, and spirants appear in all dialects, but the voicing of most consonants is positionally determined in K.km.
Intersecting the contrasting degrees of stricture are positional contrasts between points of stricture. These include labial, dental (apico- or lamino-dental), "palatal" (lamino-alveolar), "retroflex" (apico-alveolar), and velar stricture points (see Table 1). Each position changes the length of the vocal tract above the glottis, producing a distinctive harmonic filter to the consonantal noise.
Consonants typically precede vowels; but apparently in most Nuristâni dialects vowels are articulated first, followed by the formation and release of the consonant into the voicing of the already formed vowel. This transition produces stop consonants that lack aspiration, unlike the aspirated pronunciation of English initial stops.
Labials. The pronunciation of the labial consonants p, b, and m is unremarkable.
v is a spirant, labio-dental ([v]) in K.km. but bilabial ([β]) in K.ktv.; in the other languages it is labio-dental before front vowels, bilabial ([w] or [β]) elsewhere.
f is a labio-dental voiceless spirant, found only in loanwords. In conservative speech it is replaced by indigenous p.
Dentals. The dental consonants t, d, l, and n are pronounced with the tip of the tongue against the back of the upper teeth.
ć and ź are dental affricates, pronounced [ts] and [dz], respectively. The position of the tip of the tongue for these sounds marks a difference between the Northern and the Southern Nuristâni languages. Speakers of the Southern Nuristâni languages pronounce these consonants with tongue's blade hitting against the back of the upper teeth, while the tip of the tongue is anchored behind the lower front teeth of a slightly jutted jaw. Speakers of the Northern Nuristâni languages pronounce the dental affricates with the tongue's tip hitting against the back of the upper teeth.
In s and z the noisy spirantal airstream emanates from a lamino-alveolar stricture and is directed against the back of the teeth.
ź contrasts with the voiced spirant z in A., minimally with z (from loanwords) in K.km and perhaps in V. In the remaining dialects an earlier *ź lost its affrication to become z.
Lamino-Alveolars. Lamino-alveolar ("palatal") consonants are pronounced with the blade of the tongue against the upper alveolar ridge. They include the affricates č (like English "ch") and ǰ (like English "j" in "John") and the spirant š (like English "sh"). In A., V., and Tr., the voiced palatal spirant ž (like English "z" in "azure") is phonemic, contrasting with š; in K.km it is the subphonemic manifestation of š between vowels or before voiced consonants.
The status of the palatal approximant y (IPA [j]) as a phoneme distinct from i is doubtful. [j] is the normal manifestation of i before vowels
Apico-Alveolars. Apico-alveolar ("retroflex") consonants are pronounced with the tip of the tongue against the upper alveolar ridge. They are indicated by underposed dots or capital letters: ṭ, ḍ, c̣, j̣, ṣ, ẓ, ṛ, and ṇ. Also in the retroflex series are r, ř, and ň.
ṣ and ẓ are apico-alveolar spirants, similar to English "sh" and "zu", but articulated with the tongue pulled further back to produce a lower-frequency component of noise.
Similarly pronounced are the affricates c̣ and j̣, approximating English "ch" and "j," but articulated further back.
r is an upward flap, contrasting with the forward flap ṛ, which is phonemic in A., Kal., and Tr. (?), but allophonic of ṭ in K.km. A retroflex approximant ř, similar to English prevocalic r without lip-rounding, is found in all dialects except V. and in A.m, where ř has become l. ř is phonemic in K., but allophonic of either r or ṛ in A. and ṛ in Kal.
ṇ is an apico-alveolar nasal stop, similar to English "n". However, in A. and Kal. a ṇ between vowels is reduced from a nasal stop to a strongly nasalized retroflex flap, [ɽ̃], written here as ň, which is an intervocalic allophone of ṇ in Kal. and may be so in A.; but it may be minimally contrastive with ṇ in A. in word-final position (cf. A.s voň 'irrigation weir', the lone example, vs. kâṇ 'arrow').
In K.km and K.kte ň is a nasalized retroflex approximant, which phonemically contrasts with ṇ. In K.ktw the etymologically corresponding sound is ř, nasalization having been lost in that dialect.
The possibility of a retroflex lateral ḷ in A. and V. needs further investigation. In K.km there is a voiceless retroflex lateral affricate [λ̣], which is phonemically a cluster of the consonants ṭl.
Sequences of Apical Consonants. An apico-dental consonant may not precede an apico-alveolar obstruent or nasal stop. To produce such a sequence, the tongue's tip would have to slide backward from tooth to gum, an unnatural motion in the region's languages. If a normally dental consonant is followed by such an alveolar consonant, the dental consonant automatically becomes alveolar, so that the entire consonant sequence is alveolar. Because the backing of the preceding consonant is automatic, there is no need to redundantly indicate it on that consonant (see above). Thus we write sṭ ([ṣṭ]) rather than xṣṭ, nḍ ([ṇḍ]) rather than xṇḍ, tc̣ ([ṭṭṣ]) rather than xṭc̣, lṭ ([ḷṭ]) rather than xḷṭ, etc.
Note that the tongue may slide forward from alveolar to dental position, so that a retroflex consonant may be followed by a dental consonant without modification, as in K.km nâṣtʹüruk 'mosquito', šâṇtʹo˜ 'cemetery', p'eṭti 'having broken', etc.
Velars. Among the velars, K.km k has an intervocalic flapped allophone [ɣ̞]; the allophonic status of a similar voiced spirant in V. is unclear. The velar nasal ŋ (like English "ng") is phonemic in all languages.
x is a velar voiceless spirant, found only in loanwords. In conservative speech it is replaced by indigenous k.
ǧ is a velar voiced spirant, found only in loanwords. In conservative speech it is replaced by indigenous g.
q is a dorso-postvelar or pharyngeal voiceless stop, found only in loanwords. In conservative speech it is replaced by indigenous k.
Pharyngeals. ḥ is a pharyngeal voiceless spirant, found only in Arabic loanwords. In conservative speech it is not pronounced.
ˁ is a pharyngeal voiced spirant, found only in Arabic loanwords. In conservative speech it is not pronounced.
Laryngeals. h is a laryngeal voiceless spirant, found only in loanwords. In conservative speech it is not pronounced.
ˀ is a glottal stop, found only loanwords. In conservative speech it is not pronounced.
Intonation. The tonal transition from the final accented vowel to the end of an utterance carries linguistic significance. As noted above, the default intonational sequence 1 2 0 indicates simple declaration with a normally accented vowel, which is indicated by a ʹ before the vowel's symbol. If the voicing of the final accented vowel is tensed further (ʺ), a higher pitch 3 results, indicating emphatic contrast or request. Over a lengthened vowel in K.km such emphatic pitch rises, indicating supplication of the speaker.
Intonation has been closely studied only in K.km, which has at least the following transitions, or terminal intonational contours:
|1||low emphasis; preamble; wonderment|
|3 0||?||request for information|
|3 0||!||emphatic contrast|
|3 1||!?||request for explanation|
|3 3||;||more to come|
Transition. This system of transcription indicates word boundaries for clarity. Sounds within a word are produced in close transition, which implies the overlapping (coarticulation) of certain articulatory components of adjacent sounds. Across a word boundary (normally indicated by a space) sounds are normally produced in lexical transition, in which fewer types of components overlap. However, the sounds across some word boundaries may be produced in close transition, in which case the words are linked together by an underline (_) rather than separated by a space. In some cases the transition from one word to the next is even closer, with the loss of a sound adjacent to the boundary; words with such syncopated transition are linked by an equal sign (=). Words of foreign origin may be joined in accentual transition by a plus sign (+), indicating that such words are to be treated as a unit in applying final-syllable accent.
Within certain words the normal componential overlap between sounds is blocked by an open transition, symbolized by a raised dot (·). The occurrence of this so-called juncture is phonemic in at least K. and Kal. Likewise, the normal overlap of lexical transition may be broken, as indicated by a comma (,).
Symbols for the different degrees of transition in K.km appear in Table 2.
Expressive Lengthening. Expressive lengthening is a salient feature of Kâmviri
discourse. A pre-accentual consonant may be lengthened to indicate augmentation; an
accented vowel may be lengthened with rising pitch to indicate diminution, and in a couple
of words the accented vowel may be lengthened with a second stressed peak to indicate
exaggeration. Lengthened sounds are indicated by italics, and two-topped vowels by italics
and underlining, as summarized in Table 3.