Isolate A Phonology

The Phonetic inventory of Isolate A is listed below.

	Labial	Alveolar	Palatal	Velar	Glottal
Stops	p/b	t/d, ts/dz	tʃ/dʒ	k/g
Fricatives	f/v	s/z	ʃ/ʒ		h
Approximates		l, ɾ	j, ʎ	ɰ
Nasals	m	n	ɲ	ŋ

	Front	Back
High	i, y, ɪ	u
Mid	e, œ	ɔ
Low		ɑ

The symbols listed in the tables above are phonetic symbols. These will be used to transcribe Isolate A words, but not to write them. To write them, we utilize a romanization system (below) that should make the pronunciation fairly transparent. The following sounds will be written using the same letter as their phonetic symbol: p, b, f, v, m, s, z, l, k, g, t, d, i, e, and u.

Romanization and Pronunciation

This is the romanization system, which will be used to spell the language using the Roman alphabet. The full system is described in detail below. Where the romanization uses a different symbol than the IPA above, the IPA will be given:

• A, a: [ɑ] Pronounced like the “a” in “father.”
• Ä, ä: [e] Pronounced like the “a” in “gate.” Some dialects instead pronounce this [ɛ] 
  (not having merged this sound with the e below) like the “e” in “get.”
• B, b: Pronounced like the “b” in “bad.”
• C, c: [ts] Pronounced like the “ts” in “cats.”
• Cs, cs: [tʃ] Pronounced like the “ch” in “each.”
• D, d: Pronounced like the “d” in “diet.”
• Dz, dz: [dʒ] Pronounced like the “dg” in “edge.”
• E, e: Pronounced like the “a” in “gate.”
• F, f: Pronounced like the “f” in “fog.”
• G, g: Pronounced like the “g” in “goat” (never like the “g” in “genius”).
• İ, i: Pronounced like the “i” in “machine.”
• I, ɩ: [ɪ] Pronounced like the “i” in “bid.”
• J, j: Pronounced like the “y” in “yellow.”
• K, k: Pronounced like the “k” in “sky.”
• L, l: Pronounced like the “l” in “love.”
• Ly, ly: [ʎ] Pronounced like the “lli” in “million.”
• M, m: Pronounced like the “m” in “matter.”
• N, n: Pronounced like the “n” in “never”. When this letter occurs before a velar consonant <k g w>, including across word boundaries, it is pronounced [ŋ] like the “ng” in “singer” (see mutations, below).
• Ny, ny: [ɲ] Pronounced like the “ñ” in Spanish “español.
• O, o: [ɔ] Pronounced like the “augh” in “caught.”
• Ö, ö: [œ] Pronounced like the “œu” in French “sœur”, or the “ö” in German “hören.”
• P, p: Pronounced like the “p” in “spike.”
• R, r: [ɾ] Pronounced like the “r” in Spanish “pero”. Nearly identical to the “t” or “d” sound in English “matador” (pronounced quickly).
• Sz, sz: Pronounced like the “s” in “sad”. When this letter occurs before a voiced obstruent, even across word boundaries, it is pronounced [z] like the “s” in “days” (see mutations, below).
• T, t: Pronounced like the “t” in “stake”
• Tz, tz: [dz] Pronounced like the “ds” in “kids”.
• U, u: Pronounced like the “u” in “ruminate”.
• Ü, ü: Pronounced like the “u” in French “rue”, or the “ü” in German “für”.
• V, v: Pronounced like the “v” in “van”.
• W, w: [ɰ] Pronounced like the “ao in Irish “naoi.” It is produced by making a /w/ sound and keeping the lips unrounded, only making the sound with the throat.
• Z, z: [ʒ] Pronounced like the “s” n “measure”.
• Double Consonants: Doubled consonants, or geminates, occur frequently in [language]. To pronounce a doubled consonant, pronounce it twice. You might think of it as lingering over the consonant. Think of the “n” sound you pronounce in “pen knife”. It’s a longer “n” than if you pronounce the similar phrase “penny”. The same goes for the doubled consonants of Isolate A. A critical note about romanization: If a digraph (e.g. ny, sz, etc.) is doubled, only the first letter will be doubled (hence, ssz not szsz). The consonant is pronounced like a doubled consonant, though, as actual combinations such as s followed by sz are impossible.

Stress

Stress in Isolate A is lexical but typically falls on the penult. Occasionally, it falls on the final syllable. Due to the unpredictability, stress will be marked with an acute accent é, when it falls on a vowel with the umlaut (e.g. ö), it will use a doubled acute accent ő. Secondary stress is applied to every other syllable from the left of primary stress. Since it is entirely predictable based on the primary stress, secondary stress is not marked

• dzírök [ˈdʒi.ɾœk]
• dzirők [dʒi.ˈɾœk]
• felipätír [fe.ˌli.pe.ˈtiɾ]

Syllable Structure

The maximum syllable structure in Isolate A is CCVC. Any consonant may appear in the onset. The first consonant of a cluster can only be an obstruent. The second consonant of a cluster can be /ɾ/ /l/ or /n/. The only allowable nasal clusters are /pn/, /kn/, and /sn/. Alveolar stops are prohibited before /l/ e.g. */tl/. Alveolar fricatives are prohibited before /ɾ/ e.g. */ʃɾ/. The glides /ɰ/ and /j/ are only allowable coda finally when they are at the end of a word. 

Vowel Harmony

Vowels in Isolate A exhibit a vowel harmony system according to frontness and roundedness. High vowels assimilate to the fronting and rounding of the leftmost vowel in the root. The Mid vowel /e/ is transparent to the vowel harmony, that is it does not change, but does not affect the spread of the previous vowel’s attributes. Low vowels assimilate to the fronting of the leftmost vowel in the root, but are opaque to rounding changes. That is to say, a mid or low vowel blocks the further spread of rounding and spreads its own roundedness to any subsequent vowels. The vowel harmony can be described with the following chart:

	Front		Back
	Unround	Round	Unround	Round
High	i	ü	ɩ	u
Mid	e
Low Unround	ä		a
Low Round		ö		o

Mutations

Some consonants in Isolate A exhibit mutations when they occur word finally before another word. Nasal consonants /m n ɳ/ assimilate to the place of the initial consonant in the following word. Obstruents (all stops and fricatives listed above) assimilate to the voicing of the initial consonant in the following word if that initial consonant is also an obstruent.

The mutation across word boundaries does not occur when there is an audible space between words, such as might be represented by a comma or as may occur in careful speech. While the extent of mutation may vary based on the speaker, the audience, and the formality of the occasion, mutations are more likely to occur according to the following gradient:

phrase constituents < clause constituents < sentence constituents < discourse constituents

That is, elements within the same phrase are highly likely to influence each other in all circumstances, while a word at the end of one sentence is unlikely to experience mutation from the following word except in highly casual or rapid speech.

This mutation is typically reflected in the romanization and orthographic spelling, with the exception of <sz> and <n> which alternate between /s/ and /z/ and between /n/ and /ŋ/ respectively, due to the fact that these phonemes are only distinguished in these specific contexts (that is, there are no minimal pairs between these two phonemes). The typical practice for recording mutations is to only record mutations that occur between elements of the same phrase.