Unified Transcription System (UTS) for the Global Lexicostatistical Database

Unified Transcription System (UTS) for the Global Lexicostatical Database

Preliminary Notes

1. The UTS, employed in the encoding of all linguistic data in the GLD, is for the most part based on the IPA standard, but with a few significant deviations, the most important of which are listed as notes to particular tables; these deviations are explained by reasons of technical convenience, economy of symbols, and, sometimes, IPA's incompatibility with the majority of transcriptional systems in particular linguistic traditions.

2. Data encoding for the GLD is supposed to respect all of the phonological oppositions attested for the recorded language, but is not supposed to necessarily adhere to the principle of recording the exact phonetic quality for each allophone (e. g. such phenomena as retracted vowel articulation automatically imposed by an adjacent retroflex consonant may be noted, but there is no obligation for the wordlist compiler to do so; in many cases, such minutiae only clutter the transcription without being of any real help). The tables below do not, therefore, repeat all of the IPA symbolics, but mostly those parts of it that are known to have phonological validity in at least a few of the world's languages.

3. The UTS is not a closed system; further adjustments and additions are possible whenever the need arises, as the GLD accumulates more data. It is important, however, that contributors to the GLD, if faced with the need to introduce a new sound or type of articulation that the UTS fails to designate, represent it with a Unicode character or a peculiar combination of codes that is not already assigned a UTS value. Violation of this principle will lead to general confusion as well as difficulties in subjecting the data to various automatic procedures ingrained in the GLD.

4. All possible transcription signs can be found in the Starling Serif Unicode font. This is the default standard for data encoding in the GLD, and may be freely downloaded here. (It is obligatory for proper viewing of all data in Excel format).

Table 1. Principal consonants.

Place →

Labial

Coronal

Dorsal

Radical

Glottal

↓ Manner

Bilabial

Labiodental

Dental

Alveolar

Postalveolar

Retroflex

Palatal

Velar

Uvular

Pharyngeal

Epiglottal

Glottal

Nasal

n̪

Plosive

p b

p̪ b̪

t̪ d̪

t d

ʈ ɖ

ȶ ȡ

k g

q ɢ

Implosive

ɗ̪

Ejective

pʼ

t̪ʼ

tʼ

ȶʼ

kʼ

qʼ

Fricative

ɸ β

f v

θ ð

s z

š ž

ʂ ʐ

~ʆ ʑ

x ɣ

χ ʁ

ħ ʕ

ʜ ʢ

h ɦ

Affricate

pᶠ bᵛ

tᶿ dᶞ

c ʒ

č ǯ

c̢ ᶚ

ɕ ʓ

kˣ gˠ

qᵡ ɢʶ

Approximant

ʍ w

Trill

Flap

ⱱ

ɢ̆

Lateral fric.

ɬ ʫ

Lateral affr.

ƛ       ᴌ

Lateral appr.

Notes:

1. Most of the coronal affricates have been modified from IPA's original notation in order to reduce the amount of digraphs.

2. For languages with no phonological opposition between dental and alveolar stops, it is recommended to mark dental plosives as t, d, in order to avoid extra diacritics.

3. Palatal and alveo-palatal fricatives are transcribed by the exact same symbols (~ʆ, ʑ), since we are not aware of any single language that phonologically opposes the two. Since  (s with curl) is a non-Unicode symbol, it is permissible to use ʆ as a compatible replacement.

4. The palatal approximant is transcribed as y rather than IPA's j, for better transparency reasons (j is used to denote an affricate in many of the world's transcription systems).

Table 1.A. Clicks.

Efflux type	Labial	Dental	Palatal	Alveolar	Retroflex	Lateral
Zero (velar) efflux	ʘ	ǀ	ǂ	ǃ	ǃǃ	ǁ
Voiced efflux	ʘ̰ ~ ɡʘ	ǀ̰ ~ ɡǀ	ǂ̰ ~ ɡǂ	ǃ̰ ~ ɡǃ	ǃǃ̰ ~ ɡǃǃ	ǁ̰ ~ ɡǁ
Nasalized efflux	ʘ̃ ~ ɳʘ	ǀ̃ ~ ɳǀ	ǂ̃ ~ ɳǂ	ǃ̃ ~ ɳǃ	ǃǃ̃ ~ ɳǃǃ	ǁ̃ ~ ɳǁ
Glottalized efflux	ʘʼ	ǀʼ	ǂʼ	ǃʼ	ǃǃʼ	ǁʼ
Aspiration	ʘʰ	ǀʰ	ǂʰ	ǃʰ	ǃǃʰ	ǁʰ
Delayed aspiration	ʘʼʰ	ǀʼʰ	ǂʼʰ	ǃʼʰ	ǃǃʼʰ	ǁʼʰ
Preglottalized	ɂʘ̃ ~ ɂɳʘ	ɂǀ̃ ~ ɂɳǀ	ɂǂ̃ ~ ɂɳǂ	ɂǃ̃ ~ ɂɳǃ	ɂǃǃ̃ ~ ɂɳǃǃ	ɂǁ̃ ~ ɂɳǁ
Velar fricative	ʘx	ǀx	ǂx	ǃx	ǃǃx	ǁx
Velar affricate	ʘkx	ǀkx	ǂkx	ǃkx	ǃǃkx	ǁkx

Notes:

1. The retroflex click symbol is not included in IPA; it is, however, attested as an independent phoneme in North Khoisan languages and is represented here by the traditionally used symbol (double exclamation mark).

2. Transcription of the various types of click effluxes is not standardized in the IPA. For the voiced and nasalized effluxes we use R. Vossen's system (voicing = tilde below the click; nasalization = tilde above the click), but those fonts that lack these symbols may employ the alternate tradition of preceding the click symbol with a g or n, respectively (it is also advisable to use a non-standard graphic form of these letters, e. g. ɡ and ɳ, for technical recoding reasons).

3. Several of the efflux types can have slightly different manners of articulation and, accordingly, different transcriptions in various systems, such as: (a) the "zero" efflux, in most of the old sources, is perceived as a "velar" efflux, and the clicks are accordingly marked as ǀk, ǂk etc.; this norm is nowadays deemed phonetically incorrect, and most transcriptions simply supply the basic click symbol without any accompaniments; (b) the "velar fricative" and "velar affricate" effluxes, in some languages, are recognized as reflecting uvular articulation and, accordingly, transcribed as ǀχ, ǀqχ, ǂχ, ǂqχ, etc. This distinction is not, however, known to have any phonological relevance in any living or reconstructed Khoisan language.

Table 2. Vowels.

	Front		Central		Back
	Plain	Rounded	Plain	Rounded	Plain	Rounded
Close (high)	i	ü	ɨ	ʉ	ɯ	u
Near close (high)	ɪ	ʏ				ʊ (ʋ)
Close (high) mid	e	ö	ɘ	ɵ	ɤ	o
Mid			ǝ
Open (low) mid	ɛ	œ	ɜ	ɞ	ʌ	ɔ
Near open (low)	ä		ɐ
Open (low)			a		ɑ	ɒ

Notes:

1. The IPA symbols y, ø, and æ have been replaced by umlaut letters ü, ö, and ä respectively (this is particularly important for y, which is used in the UTS to denote the consonantal palatal approximant).

2. In many languages, vocalic articulation is either too blurry or too poorly described to allow the transcriber to make a single decisive choice (e. g. when choosing between ɘ and ǝ). In general, we recommend sticking to the more «common» vocalic symbols, frequently employed in phonological transcription, avoiding more rare symbols like ʉ, ʏ, ɘ, ɵ, ɜ, ɞ, ɐ, ɒ, etc., unless it is quite clear that the sounds in question make up individual phonemes or at least transparently clear and phonetically stable allophones.

Table 3. Secondary articulation features.

Syllabic	◌̩	ɹ̩ n̩	Non-syllabic	◌̯	e̯ ʊ̯
Aspirated	◌ʰ	tʰ dʰ	No audible release	◌̚	d̚
Nasal release	◌ⁿ	dⁿ	Lateral release	◌ˡ	dˡ
Voiceless phonation	◌̥	n̥ d̥	Voiced phonation	◌̬	s̬ t̬
Breathy voice	◌̤	b̤ a̤	Creaky voice	◌̰	b̰ a̰
Labialized	◌ʷ	tʷ dʷ	Palatalized	◌ʸ	tʸ dʸ
Nasalized	◌̃	ẽ z̃	Pharyngealized	◌ˤ	tˤ aˤ

Notes:

1. Several secondary features prescribed by IPA have been omitted from this table for their non-phonological nature. (In case of need, they can always be reinstated).

2. Breathy articulation for vowels can be alternately marked as aspiration (aʰ, eʰ, etc.).

Table 4. Suprasegmental features.

Primary stress (before stressed vowel)	ˈa	Secondary stress (before stressed vowel)	ˌa
Length	aː kː	Half-length	aˑ
Extra-short	ə̆	Syllable break	a.a
Minor (foot) break	\|	Major (intonation) break	‖
Global rise	↗	Global fall	↘
Extra high tone	ŋ̋ e̋ ~ e˥	Upstep	ꜛke
High tone	ŋ́ é ~ e˦	Downstep	ꜜke
Mid level tone	ŋ̄ ē ~ e˧	Rising	ŋ̌ ě
Low level tone	ŋ̀ è ~ e˨	Falling	ŋ̂ ê
Extra low tone	ŋ̏ ȅ ~ e˩	Various contour tones	ɛ᷄, ɛ᷇...

Notes:

1. It is permissible to mark tone-bearing syllables with number schemes (e. g. pa¹, maŋ², etc.), but only if the exact phonetic information on the tonal scheme cannot be found in available information sources.

2. The stress, as stated in the table, should be placed before the vowel rather than the accented syllable (i. e., patˈa, not paˈta).

3. In languages with three degrees of vowel length, extra-long vowels should be marked as long, and long vowels as half-long.