A Philosophical Grammar of Ithkuil, a Constructed Language

Ithkuil: A Philosophical Design for a Hypothetical Language

Home	5a Verb Morphology	9 Syntax
Introduction	5b Verb Morphology (continued)	10 Lexico-Semantics
1 Phonology	6 More Verb Morphology	11 The Script
2 Morpho-Phonology	7a Using Affixes	12 The Number System
3 Basic Morphology	7b Using Affixes (continued)	The Lexicon
4 Case Morphology	8 Adjuncts	Revised Ithkuil: Ilaksh

Chapter 1: Phonology

1.1 Note On Orthography and Transliteration

1.2 Phonemic Inventory

1.3 Phonological Processes and Rules

1.4 Phonotaxis

The phonology of a language essentially refers to its sound system, i.e., its systematic employment of consonants, vowels, and other vocalized phenomena such as pitch, stress (or accent), and tone in order to physically convey the meaningful content of the language itself. The phonological system of Ithkuil is detailed in the sections below.

1.1 NOTE ON ORTHOGRAPHY AND TRANSLITERATION

The native script used to represent Ithkuil is both alien and complex (employing over 3600 characters). It is explained in Chapter 11. As a result of this complexity, a system of romanization is employed throughout this grammar to allow the reader to recognize the general phonetic structure of Ithkuil words. Due to the inadequacies of the Roman alphabet in transliterating the large number of phonemes (meaningful sounds) in Ithkuil, the following diacritics are necessary for a phonemic Romanization system: the acute, grave and circumflex accents, the umlaut or dieresis, the cedilla, a superposed dot, and the apostrophe. The uses of these diacritics are explained in the sections below. Additionally, digraphs with a superscript letter h are employed to represent various aspirated consonants.

1.2 PHONEMIC INVENTORY

In furtherance of morpho-phonological efficiency (one of the language’s goals as explained in the introduction), Ithkuil must have a large phonemic inventory, specifically 65 consonants and 17 vowels, illustrated by place and manner of articulation in Table No. 1 below using a special romanized orthography. Because this is far more than most Western languages, Ithkuil has many sounds alien to speakers of English or other Western languages. Consequently, readers other than trained phoneticians (or perhaps speakers of such obscure languages as Chechen or Abkhaz) will likely have difficulty pronouncing the language accurately.

Tables 1(a) and 1(b): Phonemic Inventory

Table 1(a): Consonantal Inventory

PLOSIVES

AFFRICATES

Fricatives

Nasals

Taps/
Trills

Liquids

Approximants

plain

aspirated

ejective

plain

aspirated

ejective

un-
voiced

+voice

un-
voiced

+voice

un-
voiced

+voice

BILABIAL

p’

LABIO-VELAR

LABIO-DENTAL

DENTAL

t’

ALVEOLAR

c’

ALVEOLAR-RETROFLEX

’

POST-ALVEOLAR

’

PALATAL

’

ç’

VELAR

k’

x’

UVULAR

q’

’

PHARYNGAL

GLOTTAL

’

LATERAL

VELARIZED LATERAL

Table 1(b): Vocalic Inventory

	FRONT		CENTRAL		BACK
	unrounded	rounded	unrounded	rounded	unrounded	rounded
High	î	ÿ	_	ü	ï	û
Mid-High	i	_	_	_	_	u
Mid	ê	ø	_	_	ë	ô
Mid-Low	e	ö	_	_ _	_	o
Low	ä	_	a	_	â	_

1.2.1 Pronunciation of Consonants

The following are approximate descriptions of the consonantal sounds of Ithkuil. In addition to these descriptions, the corresponding symbol of the International Phonetic Alphabet (IPA) is provided in brackets, as well as their X-SAMPA equivalents for those readers who may be familiar with these phonetic representation systems. Sound files of actual Ithkuil words have been provided for those consonants which are uncommon or difficult to pronounce for English speakers, as well as in other cases to compare and distinguish two similar but distinct phonemes, e.g., k versus , or c vs. vs. , etc.

b	As in English. A voiced bilabial unaspirated plosive. IPA and X-SAMPA [b].
c	As in English bits. A voiceless unaspirated lamino-alveolar affricate. IPA [ts]. X-SAMPA [ t_s ]. Example sound files: ____
	No English equivalent. Similar to the sound immediately above except that the s-component of the affricate is made up of the Ithkuil alveolar retroflex s-sound (see below). This sound occurs in most Chinese languages, as well as the Basque language, where it is spelled ts. A voiceless unaspirated apico-alveolar retroflex affricate; IPA []. X-SAMPA [ t_s` ]. Example sound files:
	As in English chin but with neither the lip-rounding nor the aspiration (accompanying puff of air) that characterizes this sound in English. A voiceless unaspirated lamino-postalveolar dorso-palatal non-labialized sibilant affricate; IPA []. X-SAMPA [ t_S ]. Example sound files:
ç	Like the initial sound in English human, huge, hue, or the sound in German ich. A voiceless dorso-palatal non-grooved (slit) fricative; IPA [ç]. X-SAMPA [C]. Example sound files:
d	Similar to English, but made with the tip of the tongue against the back of the upper teeth, not the alveolar ridge (the gum ridge behind the upper teeth) as in English. An unaspirated voiced apico-dental plosive; IPA []. X-SAMPA [d_d].
	As in English this, bathe, weather. A voiced apico-interdental fricative; IPA [ð]. X-SAMPA [D].
f	As in English. A voiceless labio-dental fricative; IPA and X-SAMPA [f].
g	Always as in English gag; never as in ginger. A voiced dorso-velar unaspirated plosive; IPA and X-SAMPA [g]. Example sound file:
	No English equivalent. The voiced dorso-uvular unaspirated plosive found in Inuit and several Paleo-Siberian languages such as Chukchi and Nivkh. Similar to the hard g-sound above but made by pressing the tongue against the uvula (the little “punching bag” hanging at the back of the palate) as opposed to the soft palate. The resulting sound has a characteristic “gulped” quality. The voiced counterpart to Ithkuil q below; IPA []. X-SAMPA [G\]. Example sound files:
	No English equivalent. The voiced counterpart of x below. Can be approximated by putting the tongue in the position to pronounce English g as in gag then, without moving the tongue, trying to say ‘uh’ instead. A voiced dorso-velar fricative; IPA []. X-SAMPA [G]. Example sound files:
h	As in English hall. Note that, unlike English, this sound can occur at the end of a syllable in Ithkuil. A voiceless bi-glottal fricative; IPA and X-SAMPA [h]. Example sound files:
	No English equivalent. The voiceless radico-pharyngal fricative found in Arabic and several Northeast Caucasian languages. A deep raspy sound produced by pronouncing an h-sound while constricting the pharynx and trying to press the root of the tongue against the back wall of the pharynx. IPA []. X-SAMPA [X\]. Example sound files:
j	As in English judge but without the lip-rounding that accompanies the English sound. An unaspirated non-labialized voiced lamino-postalveolar dorso-palatal affricate; IPA []. X-SAMPA [d_Z]. Example sound files:
	No English equivalent. The voiced counterpart of Ithkuil below. Found in Hungarian and Czech where it is spelled gy and d´ respectively. Sort of like a g+y sound as in English big year pronounced rapidly, with the g-sound pronounced with the tongue against the hard palate as opposed to the soft palate. A voiced dorso-palatal unaspirated plosive. IPA []. X-SAMPA [J\]. Example sound files: ‘day (24-hour period)’
k	Similar to English k but without aspiration. Like the k-sound of the Romance languages, e.g., Spanish or Italian casa. A voiceless unaspirated dorso-velar plosive; IPA and X-SAMPA [k]. Example sound files:
	No English equivalent. Like a k-sound but unaspirated (i.e., without any accompanying puff of air) and produced farther forward in the mouth by pressing the tongue to the hard palate, not the soft palate as with English k. The result should sound somewhat like a k+y as in backyard when spoken rapidly. A voiceless unaspirated dorso-palatal plosive; IPA []. X-SAMPA [c]. Example sound files:
l	The “light” l-sound of the Romance languages, or as in British English leader; not the “dark” (velarized) l-sound of American English lull. A voiced apico-dental dorso-bilateral liquid continuant; IPA and X-SAMPA [l]. Example sound files:
	The “dark” (velarized) l-sound of American English lull, not the “light” l-sound of the Romance languages or British English leader. A voiced velarized apico-dental dorso-bilateral liquid continuant; IPA []. X-SAMPA [5]. Example sound files:
	No English equivalent. The voiceless dorso-bilateral fricative as found in Welsh llan. Can best be approximated by putting the tongue in position as if to say an l-sound, and while holding the position, make a forceful h-sound instead; IPA []. X-SAMPA [K]. Example sound files:
m	As in English. A voiced bilabial nasal continuant obstruent; IPA and X-SAMPA [m].
n	Similar to English, but made with the tip of the tongue on the back of the upper teeth as in the Romance languages, not the alveolar ridge as in English. A voiced apico-dental nasal continuant obstruent; IPA []. X-SAMPA [n_d].
	The English ng-sound as in song or ringer; NEVER as in finger. A voiced dorso-velar nasal continuant obstruent;. IPA []. X-SAMPA [N].
p	Similar to English, but without aspiration. As in the Romance languages. A voiceless unaspirated bilabial plosive; IPA and X-SAMPA [p].
q	No English equivalent. The voiceless dorso-uvular unaspirated plosive found in Arabic, Inuit and many American Indian and Caucasian languages. Similar to an unaspirated k-sound but made by pressing the tongue against the uvula (the little “punching bag” hanging at the back of the palate) as opposed to the soft palate. IPA and X-SAMPA [q]. Example sound file:
	This sound is a combination of a heavily aspirated and palatal Ithkuil (see above) released laterally (i.e., along the two sides of the tongue) into a voiceless lateral fricative (IPA []). This sound occurs in some North American Indian and North Caucasian languages. An aspirated voiceless palatal-lateral affricate. IPA []. X-SAMPA [c_K_h]. Example sound files:
r	This sound is a single flap of the tongue tip as in Spanish caro or pero. When geminated (doubled) it becomes a trill as in Spanish carro or perro. A voiced apico-alveolar retroflex flap/trill; IPA [], [r]. X-SAMPA [4], [r]. Example sound files:
	Pronounced similarly to the English retroflex r-sound except that there is no lip-rounding and the tongue is in direct contact with the alveolar gum ridge behind the upper teeth, giving the sound an l-coloring. A voiced apico-alveolar retroflex liquid continuant; the closest IPA symbol is [] (a voiced rhoticized retroflex lateral approximant). (No X-SAMPA equivalent). Example sound files:
	No English equivalent. Like the throaty or “gargled” r-sound found in colloquial French and German. A voiced dorso-uvular approximant (non-trilled); IPA []. X-SAMPA [R]. Example sound files:
s	As in English sister. A voiceless lamino-alveolar grooved sibilant fricative; IPA and X-SAMPA [s]. Example sound file:
	No English equivalent. The retroflex s-sound heard in Castilian (but not Latin American) Spanish. Also found in Basque, Catalan, and most Chinese languages. Similar to an English s-sound, but with the blade of the tongue curved convexly away from the roof of the mouth, as with an English r-sound, so that the tongue curls back and touches the alveolar ridge. The result should sound halfway between an English s and sh. A voiceless apico-alveolar retroflex grooved sibilant fricative; IPA []. X-SAMPA [ s`]. Example sound file:
š	As in English shoeshine but without the lip-rounding of the English sound. A voiceless non-labialized lamino-postalveolar dorso-palatal grooved sibilant fricative; IPA []. X-SAMPA [S].Example sound file:
t	Similar to English but without aspiration and with the tongue-tip against the back of the upper teeth, not against the alveolar ridge. As in the Romance languages. A voiceless apico-dental unaspirated plosive; IPA []. X-SAMPA [t_d].
	As in English thin, bath. A voiceless apico-interdental fricative; IPA []. X-SAMPA [T].
v	As in English. A voiced labio-dental fricative; IPA and X-SAMPA [v].
w	As in English well, worry. A voiced labio-velar (i.e., labialized dorso-velar) glide (or approximant); IPA and X-SAMPA [w].
x	No English equivalent. The smooth voiceless dorso-velar fricative found in Russian (spelled x in Cyrillic) and in Latin American (but not Castilian) Spanish j. Can be approximated by putting one's tongue in position as if to pronounce a k-sound, and while holding the tongue in this position, breathing an h-sound instead. Should be distinguished from below; IPA and X-SAMPA [x]. Example sound file:
	No English equivalent. The rough voiceless dorso-uvular fricative (or trill) found in German ach. Can be approximated by means of a dry gargle without vocal chord vibration. IPA []. X-SAMPA [X]. Example sound file:
y	As in English yet, yam. A voiced dorso-palatal glide (or approximant); IPA and X-SAMPA [ j ]. Example sound file:
	No English equivalent. The voiced counterpart to ç above. This is the “intensive” y-sound heard in Castilian and some Latin American varieties of Spanish. Like the sound in English yet, yam but with the blade of the tongue held more closely to the hard palate, so that audible friction occurs. A voiced dorso-palatal fricative; IPA []. X-SAMPA [j\]. Example sound file:
z	As in English zoo, wizard. A voiced lamino-alveolar grooved fricative; IPA and X-SAMPA [z]. Example sound file:
	No English equivalent. The voiced counterpart to Ithkuil above. A voiced apico-alveolar retroflex grooved sibilant fricative. IPA []. X-SAMPA [ z`]. Example sound file:
	The voiced counterpart of š above. Similar to the sound in English pleasure or leisure, but without lip-rounding. A voiced lamino-alveolar dorso-palatal grooved sibilant fricative; IPA []. X-SAMPA [Z]. Example sound file:
	Like an English d+z sound, as in roads, adze. A voiced lamino-alveolar affricate; IPA [dz]. X-SAMPA [d_z]. Example sound file:
	No English equivalent. The voiced counterpart to above. A voiced apico-alveolar retroflex affricate; IPA []. X-SAMPA [d_z`]. Example sound file:
’	This sound is the glottal stop heard between the two vowels in English oh-oh or as the sound heard in most American English speakers’ pronunciation of the word fattening. This sound is very common in other languages such as Hawaiian, Arabic, Hebrew, etc. A voiceless bi-glottal stop; IPA []. Example sound files:

1.2.1.1 Aspirated Consonants: The consonants are all unaspirated, i.e., without the accompanying puff of air characteristic of English voiceless stops and affricates. In Ithkuil each of these has an aspirated counterpart, pronounced like the unaspirated version but with a distinct expulsion of air, more so than in English. These aspirated counterparts are written with a following superscript h; thus: Example sound files:

1.2.1.2 Ejective Consonants: The same eight consonants immediately above also have ejective counterparts, which do not exist in any major Western language, but are found in languages such as Armenian, Amharic, Georgian, most of the Caucasian languages, and many American Indian languages. Ejectives (also called glottalized consonants) are consonants accompanied by simultaneous closure and sudden release of the glottis (vocal chords), which gives the sound a distinct “popped” or explosive quality. Ejectives are indicated by an apostrophe following the consonant, thus: . Note: the lateral affricate has an allophone (alternate pronunciation) as an ejective as well. Example sound files:

In addition to the above ejectives, there are three additional ejective consonants, all of them affricates, ç’, x’, and ’, whose rather difficult and exotic pronunciations are described below.

ç’	This sound is a combination of Ithkuil + Ithkuil ç (the voiceless palatal slit fricative described above), the combination then ejectivized. There is no English approximation. When pronounced properly, it should sound like a sudden high-pitched explosive hissing. IPA [’]. X-SAMPA [c_C_>]. Example sound file:
x’	This sound is a combination of Ithkuil k + Ithkuil x (the voiceless velar fricative described above), the combination then ejectivized. There is no English approximation. When pronounced properly, it should be a harsh, sudden, emphatic k-sound, accompanied by a high-pitched sound of static, as if someone is verbally imitating the sound of a glass breaking. IPA [kx’]. X-SAMPA [k_x_>]. Example sound file:
’	This sound is a combination of Ithkuil q + Ithkuil (the voiceless uvular fricative described above), the combination then ejectivized. There is no English approximation. When pronounced properly, it sounds like a very harsh, explosive, strangled choking sound. IPA [q’]. X-SAMPA [q_X_>]. This sound is found in a few obscure Northeast Caucasian languages (e.g., Dido, Archi). Example sound file:

1.2.1.3 Syllabic consonants. Six consonants, , can be pronounced as full syllables in absence of a vowel. The phenomenon of syllabic consonants is fairly common and occurs in colloquial English expressions such as ‘hmm’ (as when pondering a thought), ‘mm-hmm’ (an expression of approval or agreement), as well as with the consonants n and l as in the second syllable of words like button and little. Syllabic nasals are also found in Navajo, as in the words nda ‘no’, and ndíghílii ‘sunflower.’ In Ithkuil, these six syllabic consonants can appear as word-initial syllables preceding a consonant as in However, they also occur in special geminate (i.e., doubled) clusters where the second “half” of the geminated cluster is pronounced as a separate syllable. Examples: .

1.2.2 Pronunciation of Vowels

There are 17 vowels, all of which are pure sounds, not glided into diphthongs as in English. These include the five primary vowels a, e, i, o, u. The vowel a is phonetically an unrounded central low vowel, IPA [a], as in Spanish or Italian. The vowels e and o are similar to the vowels in American English let and short, phonetically IPA [] and [], although both sounds are actually somewhat higher, being between low-mid and mid in height. The vowels i and u are lower than in Romance languages: i is about halfway between the vowels in English pit and machine; u is halfway between English cook and kook. Sound files: a e i o u

The vowel â is pronounced as in Western U.S. all, IPA []; ê and ô are mid-height, IPA [e] and [o] as in Spanish estos or in French poésie; î and û are high (i.e., fully closed) as in Spanish or Italian i and u, IPA [i] and [u]. Example sound files: â ê î ô û

The vowel ä is slightly higher than the vowel in American English sat but not as high as in set, IPA [æ^]. The vowel ö is the rounded equivalent of e, that is, the vowel in French boeuf or German könnte, IPA [œ]. The vowel ë is pronounced somewhat like the vowel in American English cut or nut, although, more exactly, it is the Ithkuil vowel ô but without rounding of the lips, a vowel which occurs in Estonian, IPA []. The vowel ï is pronounced as an unrounded û, an obscure vowel found in Turkish and Japanese, IPA . The vowel ü is pronounced as a high central rounded vowel, as found in Norwegian hus or the Highland Scottish pronunciation of English book or good, IPA []. The vowel ø is the rounded equivalent of ê, as in French feu or German schön, IPA [ø]; the vowel ÿ represents the front rounded vowel of French du and German über, IPA [y]. Example sound files: ä ë ï ö ø ü ÿ

1.2.3 Allophonic Distinctions

Allophonic distinctions are the phonetic variances in the pronunciation of a particular phoneme depending on the phonetic environment in which that phoneme occurs. These variances, while audible to a trained linguist, are often indistinguishable to lay native speakers of a given language, in that these allophonic variances do not change the meaning of a word and thus play no functional role in the language. As an example, compare the two t-sounds in the English words top and stop. The former is aspirated (i.e., accompanied by a distinct puff of air), while the latter is unaspirated, giving the two sounds a different phonetic quality. However, because consonant aspiration does not function phonemically in English, the difference in the two t-sounds is unnoticeable to most native speakers of English, even though it would be highly noticeable to speakers of languages where consonant aspiration is phonemically relevant (e.g., Hindi and many other Indic languages).

Although such allophonic distinctions are arbitrary within a given language, they are not random; rather, their patterns are completely regular and predictable for any given language (as is true for consonant aspiration in English). Failure to follow the rules for allophonic distinctions when learning a foreign language will result in the speaker having a noticeable “foreign accent” to native speakers of the language (as do most French, Italians, and Spanish-speakers when trying to pronounce English “top” without aspirating the initial t-sound, due to the lack of consonant aspiration in Romance languages.)

The particular phonetic variants of a particular phoneme are known as allophones. The significant allophonic distinctions for Ithkuil are as follows:

h	This phoneme is a simple (bi-)glottal fricative as in English head when in a syllable-initial position and in word-final position. However, at the end of a syllable when preceded by a vowel and followed by another consonant, this sound approaches a voiceless bilabial fricative, similar to the bilabial sound of Japanese h as pronounced before u (usually transliterated into Roman orthography as f). IPA [].
hh	The geminated version of Ithkuil h is pronounced as a “bi-dental” fricative, in that the jaw is completely closed and the upper and lower teeth are in near-contact along their entire length. The resulting sound is somewhat similar in timbre to both a voiceless interdental fricative (as in English thin) as well as the English f-sound, however there is absolutely no contact by the tongue with the teeth or gums when pronouncing this sound. No IPA equivalent. Example sound files:
	This aspirated lateral affricate is in free variation with (i.e., may be alternately pronounced as) its non-aspirated ejective counterpart (IPA []); in word-initial position it is more common to pronounce it ejectively. Example sound files:
	In normal speech, this phoneme is pronounced as a voiced dorso-uvular approximant (non-trilled) continuant, similar to the throaty r-sound found in colloquial French and German, IPA []. However, in emphatic articulation or hyper-enunciated speech, this sound becomes a voiced dorso-uvular trill, IPA [].
w	Normally pronounced as a voiced labio-velar approximant (i.e., labialized dorso-velar approximant) as in English wet (IPA [w]), when followed by the Ithkuil vowel û, this sound, takes on even greater lip-rounding to become a voiced labio-velar fricative (i.e., labialized dorso-velar fricative) (IPA []).
w’	In this word-initial combination of voiced labio-velar approximant followed by a glottal stop is found, the w is pronounced followed by a brief high central unrounded vowel, Ithkuil ï, followed by the glottal stop (IPA []).
y’	Similarly to the combination above, this word-initial combination has its voiced dorso-palatal approximant followed by a brief high central unrounded vowel, followed by the glottal stop (IPA []).
bm, dn, km, kn, pm, tn
	When in word-initial position, the first consonant of these conjuncts is pronounced with nasal rather than oral release. To achieve this, place the tongue and/or lips in position to pronounce the first consonant, initiate the airstream from the lungs to pronounce it, but instead of releasing the sound, and without moving the tongue or lips, pronounce the second nasal consonant instead. IPA [].
hl, hm, hn, h, hr, h, hw
	Each of these consonant conjuncts, when word-initial, or syllable-initial following another consonant, are not pronounced as separate consonants, but rather as unvoiced counterparts to the liquid or nasal consonant that forms the second member of the conjunct, i.e., IPA To approximate these sounds, place the mouth in the position to pronounce an Ithkuil l, m, n, n, r, , or w, and without moving the tongue or lips, breath a clear h-sound instead.

1.3 PHONOLOGICAL PROCESSES AND RULES

Besides the actual inventory of consonant and vowels, all languages have phonological processes which affect how those consonants or vowels are combined and phonetically articulated. Through these phonological processes, the possible number of word-forming syllables in the language is expanded. Ithkuil productively utilizes consonantal gemination, shifts in syllabic stress, and tone (pitch intonation) to achieve these ends.

1.3.1 Gemination

Gemination refers to the audible “doubling” in length of a particular sound, usually in reference to consonants. While gemination does not occur in English on true phonological grounds, it does occur on morpho-phonological grounds, as seen in the difference in pronunciation of the phrase ‘a natural’ versus ‘unnatural.’ There are many languages, however, where phonologically-based gemination is an intrinsic component of the phonology (e.g., Italian, Japanese, Finnish).

In Ithkuil, most consonants can be geminated. Also noteworthy is that gemination of certain consonants is allowed in both word-initial and word-final position. The following are the specific rules for consonant gemination in Ithkuil:

1.3.1.1 Intervocalic Gemination. All consonants are capable of intervocalic gemination (i.e., when between two vowels) except for y and w.

1.3.1.2 Word-Initial and Word-Final Gemination. The following consonants, in addition to being capable of intervocalic gemination, are also capable of being geminated in both word-initial and word-final position:

The consonant can be geminated in word-initial position.

1.3.1.3 Pronunciation of Geminated Consonants. Consonants which are continuants (i.e., able to be sounded for an indefinite duration), specifically are simply pronounced for twice as long in duration when geminated. Geminated r is pronounced as a rapid apico-alveolar trill like rr in Spanish or Italian. Example sound files:

The non-aspirated plosive consonants b, d, g, , , k, , p, q, and t, when geminated, are momentarily held, then released, much like the two d-sounds in the English phrase bad dog when spoken rapidly. Example sound file:

The aspirated plosives are likewise held momentarily before release, the aspiration occurring upon release, much like the two t-sounds in the English phrase hot tar when spoken rapidly. Similarly, the ejective plosives k’, ’, p’, q’, and t’ are also held momentarily before release, the glottalic ejectivization occurring upon release. Example sound files:

The pronunciation of affricates () when geminated depends on whether or not they are in word-final position. If not in word-final position, they are pronounced by momentarily holding the initial stop (plosive) component of the affricate before releasing it into the fricative or sibilant portion. Aspiration or ejectivization, if present, occurs during release of the plosive into the sibilant or fricative component, e.g., is pronounced as IPA [ttš]. For those affricates which can appear as word-final geminates (), geminated pronunciation in word-final position is achieved by simply lengthening the sibilant continuant portion of the affricate (i.e., the second sound of each affricate). Thus, in word-final position is pronounced as IPA [tšš], as IPA []. Example sound files:

1.3.1.4 Romanized Orthography of Geminates. Single character consonants are simply written double when geminated, e.g., bb, dd, nn, šš. Aspirated digraph-consonants have the first letter of the digraph written doubled followed by a single superscript h, e.g., Geminated ejective consonants are likewise written with the initial character doubled followed by a single apostrophe, e.g., Exceptions to this rule exist for the ejective affricates ç’, x’, and ’. Because the non-apostrophed forms of these three characters do not correspond to non-ejectivized versions of the apostrophed form, these geminates are written as

1.3.2 Pitch and Tone

Ithkuil is a tone language like Chinese, Vietnamese, and other Southeast Asian languages as well as most of the sub-Saharan African languages and some American Indian languages. This means that pitch or tone of voice is used to convey grammatical information, unlike Western languages which use tone and pitch changes “supra-segmentally” to mark various morpho-semantic features. For example, in English rising intonation of the voice signals a question, while other specific pitch contours signify emphasis, disgust, irony, and other attitudes. Ithkuil marks such features morphologically, i.e., within the words themselves (such as with affixes or variances in mood categories).

There are five tones used in Ithkuil, one of which, mid-low tone, is considered morpho-phonologically neutral. The other four tones are falling, high, broken (i.e., mid to low to mid-low), and rising (i.e., mid-low to mid-high) and are considered functionally significant. Each word carries one significant tone, pronounced beginning with the stressed syllable and continuously carried through any following syllables until the end of the word. Unstressed syllables prior to the stressed syllable have neutral mid-low tone. Therefore, the function of mid-low tone is solely to indicate the start of a new word since any preceding word must end in a tone other than mid-low. Like most tone languages, the tones do not correspond to any exact pitch, but are relative for each individual speaker and utterance. The relative pitch of the tones is illustrated below:


(mid-low)	falling	high	broken	rising

The four significant tones are indicated in the Roman transliteration by small superscript symbols at the end of each word, as follows: falling tone is unmarked, high tone is marked by a superscript hyphen (or macron), broken tone by a superscript backslash, and rising tone by a superscript forward slash. This is illustrated in the example words/sound files below.

[falling tone]
[high tone]
[broken tone]
[rising tone]

1.3.3 Syllabic Stress

Stress normally falls on the penultimate (next-to-last) syllable but can shift to either the ultimate (last) syllable of a word, to the antepenultimate (third-from-last) syllable, and occasionally to the preantepenultimate (fourth-from-last) syllable, as determined by morphological (grammatical) considerations.

When transliterating Ithkuil into Roman characters, the number of vowels in Ithkuil requires the use of diacritics due to the limited number of Roman vowel characters. As a result, the orthographic representation of stress using the system of romanization becomes complicated. It is explained as follows:

1)	All monosyllabic words are unmarked for stress.
2)	The grave accent ( ` ) designates an unstressed syllable when stress would otherwise be misinterpreted, e.g., in distinguishing monosyllabic diphthongs such as au and ei from dissyllabic vowel conjuncts aù and eì.
3)	Penultimate stress. Polysyllabic words having penultimate stress are unmarked for stress, except for those containing the dissyllabic phonemes ì or ù as the penultimate syllable, which, if stressed, take an acute accent, e.g., the word iskoùt (stress on the o), if adding the syllable -ma, becomes iskoútma (stress on the u).
4)	Ultimate stress. Polysyllabic words which have ultimate stress indicate this in one of the following ways:
	the unmarked vowels a, e, i, o, u, and the marked disyllabic vowels ì and ù take an acute accent, e.g., á, é, í, ó, ú. if the vowel to be stressed already carries a diacritic mark (other than the grave accent) and the vowel, diphthong, or syllabic liquid or nasal in the penultimate syllable does not, then this penultimate syllable takes a grave accent, e.g., ròihnäl. If the word has at least three syllables and the vowels or diphthongs in the last two syllables carry diacritics, then a grave accent over the antepenultimate (third-from-last) syllable implies ultimate stress (as the grave accent would be unnecessary if the word carried penultimate stress), e.g. òspätlök. if the vowels (or diphthongs) in both the ultimate and penultimate syllables already have diacritic marks (other than the grave accent) then the stressed vowel is written double, e.g., ksûtpäär.
5)	Antepenultimate stress. To show antepenultimate stress (third syllable from the end):
	if the stressed vowel (or diphthong or syllabic liquid or nasal consonant) does not carry a diacritic (other than the grave accent), it takes the acute accent, e.g., áksiyor if the stressed vowel already has a diacritic (other than the grave accent), then use the grave accent plus the lack of acute accent on the ultimate syllable or other non-diacriticized vowels to indicate stress, e.g., ëitlàrrun, ôrümzìl. (Note that the use of the grave accent over the i in ôrümzìl would be unnecessary if the word carried penultimate stress, therefore its presence implies antepenultimate stress). if existing diacritics prevent clear indication using the above rules, then the stressed vowel is written doubled, e.g., öömoläk.
6)	Preantepenultimate stress. To show preantepenultimate stress (i.e., fourth syllable from the end) apply the same rules as for antepenultimate stress above, but to the preantepenultimate syllable.

Note that the acute accent on the stressed second member of the bi-syllabic geminate clusters distinguishes them from their standard geminate counterparts .

1.4 PHONOTAXIS

In addition to phonological processes such as gemination, stress-shifting, and tone, all languages employ their own individual and arbitrary rules as to what combinations of consonants and vowels are permissible in a syllable or word. This concept is called phonotaxis and such rules are known as phonotactical rules. These rules, peculiar to each language, explain why sprelch could be a hypothetical word in English, while znatk could not be, even though znatk is as easily pronounced by a linguist as sprelch. Rules governing syllable structure, diphthong formation, and overall phonetic euphony are all part of phonotaxis

Equally important are the optional rules each language employs to achieve euphony and greater ease of pronunciation, known as phonaesthetics or phonaesthetic rules. Together, phonotaxis and phonaesthetics are greatly responsible for the phonetic “character” or subjective “sound” of a given language. The phonotactic and phonaesthetic rules for Ithkuil are described in the sections below.

1.4.1 Syllable Structure

The permissible syllable structure depends on whether the syllable forms a monosyllabic word, is a word-initial syllable, a word-final syllable, or is word-medial (i.e., between two other syllables). These structures are shown in Table 2 below, where (C) represents an optional consonant and V represents a mandatory vowel or diphthong.

Table 2: Syllabic Structure

Syllable Type	Structure	Examples
Consonantal Word	C(C)(C) where final consonant is a nasal, liquid, or voiceless fricative continuant	s, h,, ll, mm, pçç
Monosyllabic	(C)(C)(C)V(C)(C)(C)	a, ui, öt, isk, du, tuil, kleb, tliqs, pskarn, xxoršt
Word-initial	(C)(C)(C)V(C)- or l, , , m, n,	uran, tahin, ui’wá, prinu, klatma, xmoiskra, kstollap, ltuirbis, mpeilt’um
Word-medial	-(C)(C)V(C)(C)- or l, , , m, n,	kialun, ruentik, isteixlam, ïkspûzqai,
Word-final	-(C)(C)V(C)(C)(C)- or l, , , m, n,	lua, antoi, tial, eifqés, poxšurn, ultrönn,

The following rules apply:

A single non-syllabic intervocalic consonant is syllabically part of the vowel following, not preceding.
An intervocalic bi-consonantal conjunct is considered dissyllabic (i.e., each consonant is part of a different syllable), even if the conjunct is a geminate consonant.
If a triple-consonant conjunct occurs in the middle of a word, the first consonant of the three must be in a separate syllable from the last consonant of the three. The middle consonant will be in the same syllable as whichever of the two surrounding consonants it forms a permissible word-initial or word-final cluster. If it can form permissible clusters with either surrounding consonant, it will be considered part of the syllable with which it is morphemically associated. Section 1.4.4 below details which specific consonants can form such clusters.

1.4.2 Diphthongs

An Ithkuil syllable may contain one diphthong (a combination of two vowels pronounced together as one syllable). All Ithkuil diphthongs are “falling” diphthongs, i.e., the first vowel of the diphthong receives the primary articulation while the second becomes semi-vocalic (sometimes referred to as semi-consonantal or an “off-glide”). There are 24 diphthongs in Ithkuil, described as follows:

ai	Pronounced as in Spanish or Italian; like English i in white or ice.
äi	No English Equivalent. A combination of the Ithkuil vowel ä plus an English y-sound.
ei	As in Spanish or Italian; like English ai in rain or ei in rein.
ëi	No standard English equivalent; a combination of the vowel sound in American English rut + an English y-sound. Somewhat like a rural British dialectal pronunciation of the i in ice.
oi	As in Spanish or Italian; as in English boy or voice.
öi	No English Equivalent. A combination of the Ithkuil vowel ö plus an English y-sound. Somewhat like the French word oeil.
ui	As in Spanish or Italian; no English equivalent. A combination of Spanish or Italian “pure” u (i.e., without the glide into -w as in English rude) + an English y-sound. The speaker should avoid allowing this diphthong to become a “rising” diphthong where the u-sound is reduced to a w- (the result sounding like English wee).
au	As in Spanish or Italian; like English ou in loud or ow in cow.
äu	No English Equivalent. A combination of the Ithkuil vowel ä plus an English w-sound.
eu	As in Spanish or Italian; no English equivalent. A combination of “pure” e (i.e., without the glide into –y as in English they) + an English w-sound.
ëu	No standard English equivalent; a combination of the vowel sound in American English rut + an English w-sound. Somewhat like a rural British dialectal pronunciation of the word oh!.
iu	No English equivalent; a combination of the Ithkuil vowel i (which is more open than the Spanish or Italian “pure” i (i.e., closer to English i in bit) + an English w-sound.
ou	Like the o + w-glide of English road or mode. Also as in Brazilian Portuguese roupa.
öu	No English Equivalent. A combination of the Ithkuil vowel ö plus an English w-sound. Somewhat like an exaggerated upper class British pronunciation of the word oh!.
_aï, eï, ëï, iï, oï, uï
	Each of these is a combination of a vowel with the back central vowel ï. No English equivalents, although the sound can be approximated by pronouncing the first vowel followed by the velarized “dark” l-sound of American English lull but without touching the tip of the tongue to the gum ridge behind the upper teeth; the tongue tip should remain low instead. The resulting diphthongs should sound somewhat like English all, ell (as in bell), ull (as in dull), eel, ole (as in pole), and ool (as in tool), as pronounced by someone speaking in an indistinct, slurred voice. For those familiar with IPA, these diphthongs may be represented as
_äï, öï, üï
	As with the series of diphthongs immediately above, these are combinations of the Ithkuil vowels ä, ö, and ü with the back central vowel ï. IPA
ae	No English equivalent. A combination of Ithkuil a + ê. Sounds like a more “open” version of Ithkuil ai. Tolkien’s Sindarin language also has this diphthong.

All other combinations of vowels are dissyllabic, i.e., are pronounced as two separate syllables. Care should be taken to avoid collapsing the many two-vowel combinations beginning with u- and i- into “rising” diphthongs beginning with a w-sound or y-sound. This is especially important when the second vowel of these combinations receives the syllabic stress.

1.4.3 Constraints on Vowels

All the vowels, diphthongs, and dissyllabic vowel conjuncts previously mentioned can occur in any syllable in any position with the following exceptions:

In a word of two syllables or more, the only vowels, diphthongs, or vowel conjuncts that may occur in word-final position are: -a, -â, -ï, -u, -ü, û, -ia, -ua and diphthongs ending in i, u, or ï. Note that this constraint does not apply to monosyllabic words.
No more than two vowels may occur in conjunction within a word, i.e., combinations of three vowels, even if two represent a diphthong, are not permissible.

1.4.4 Constraints on Consonants

The consonants w, y, and cannot appear in syllable-final position and must always be followed by a vowel, or in the case of w and y, by a glottal stop plus a vowel (e.g., w’+vowel).
The glottal stop (’) does not occur word-initially except as a juncture feature (i.e., where two separate words come together). Specifically, when a word ends in a vowel and the next word begins with a vowel, Ithkuil phonetically separates the words by inserting a glottal stop at the beginning of the second word so that the word-final vowel of the first word does not combine with the word-initial vowel of the second. This serves to keep the two words separate and distinct. This glottal stop is unmarked in both the Roman transliteration and in the Ithkuil script. Elsewhere, the glottal stop may appear following w or y if preceding a vowel, or may follow a vowel.
No more than four consonants can occur in conjunction. Such four-consonant conjuncts can only occur intervocalically (or across word boundaries), and the first two and last two consonants must each form permissible syllable-final and syllable initial pairs respectively, or alternately, the first three of the four consonants must form a permissible syllable-final conjunct or the last three of the four consonants must form a permissible syllable-initial conjunct.
For intervocalic three-consonant conjuncts either the first consonant or the last consonant of the conjunct must be in a separate syllable than the other two consonants. If the first consonant is dissyllabic from the other two, then those other two consonants must be a permissible syllable-initial conjunct. If the third consonant is dissyllabic from the first two, then the first two must be a permissible syllable-final conjunct.
For word-initial and word-final triple consonant conjuncts, these are generally permissible if the both the first two consonants are permissible as a conjunct by themselves and the last two consonants of the conjunct are permissible as a conjunct by themselves.

Table 3 below shows the permissible structures for word-initial consonant conjuncts, while Table 4 shows the permissible structures for consonant conjuncts in word-final position. Note that not all the possible conjuncts are shown, merely examples of each type of combination. There are many combinations of consonants permitted in Ithkuil which would never be allowed as conjuncts in Western languages.

Table 3: Permissible Structure of Initial Consonant Conjuncts

Table 4: Permissible Structure of Word-Final Consonant Conjuncts

1.4.5 Phonaesthetic Rules

Phonaesthetic rules refer to the generally applied rules and preferences peculiar to each language for structuring the patterns of phonemes for purposes of phonological euphony. The following such principles apply to Ithkuil.

While “open” syllables (syllables ending in a single vowel) are tolerated in Ithkuil, “closed” syllables (i.e., ending in a consonant) are preferred, especially for the stressed syllable of a word. Ithkuil morphology often allows for more than one way to structure the various morphemes of a particular word, especially words of three syllables or more. (Morphemes are the individual meaningful parts of a word, e.g., the root, any affix, any mutation, or any other phonological manifestation of a grammatical category.) It is common to restructure words via morpho-phonological manipulation so that they contain closed syllables, particularly in the stressed syllable. This rule is not often applied to two-syllable words, which have fewer options in their structure. The rule is inapplicable to monosyllabic words, where open syllables are common.
Words of six syllables or more are considered euphonically undesirable, and it is common to morphologically restructure a word to reduce the number of syllables to five or less. This morpho-phonological restructuring is accomplished by alternating adjunct and suffix forms (as explained in Chapters 7 and 8).
In words having multiple affixes, it is common to use one of the affixes in an adjunct form instead, if this will reduce the number of syllables.

Proceed to Chapter 2: Morpho-Phonology >>

Home	5a Verb Morphology	9 Syntax
Introduction	5b Verb Morphology (continued)	10 Lexico-Semantics
1 Phonology	6 More Verb Morphology	11 The Script
2 Morpho-Phonology	7a Using Affixes	12 The Number System
3 Basic Morphology	7b Using Affixes (continued)	The Lexicon
4 Case Morphology	8 Adjuncts	Revised Ithkuil: Ilaksh