by Don Sena
This material is adapted from: Noss, Richard. B. (1964). Thai reference grammar. Washington: Foreign Service Institute, Dept. of State.

Concept of the Phoneme:

Most of us probably realize, at this stage of our learning, that Thai syllables carry a supra-segmental feature called the syllable tone, a distinctive feature in that minimal pairs exist to show that a mere difference in tone will be accompanied by a complete difference in meaning (or no meaning). For example, consider ข้าว  /khâaw/ versus ข่าว  /khàaw/, ท่อ  /thɔ̂ɔ/ versus ท้อ  /thɔ́ɔ/, and พี่  /phîi/ versus ผี  /phǐi/. There are at least five (in reality, six) significant tonal contrasts. Because of this significance, each of these contrasts defines what's known as a phoneme.

Variants of any one phoneme (whether vowel, consonant or tone) are called allophones. A phoneme, then, is actually set of one or more allophones. An allophone, in turn, is a phonetic segment, the overt expression of a phoneme in a particular environment. The allophones of a particular phoneme constitute an “allophonic set.”

In English, the phoneme /p/ has allophones [ph] and [p]. Thus, /p/ ➔ [ph], [p], in which [p] is an unaspirated bilabial voiceless stop and [ph] is the corresponding aspirated stop. We say that /p/ ➔ (is rendered as) [p] in clusters after /s/ and as [ph] otherwise. The allophone [p] occurs in ‘spill’ and [ph] in ‘pill.’ [p] and[ph] are said to be in “complementary distribution,” where one occurs the other does not, and vice-versa. [p] and [ph] make up an allophonic set in English.

In Thai, of course, /p/ and /ph/ are phonemically distinct: ป้า  /pâa/ and ผ้า  /phâa/, for instance. Accordingly, we see that /p/ and /ph/ are not allophones of the same phoneme in Thai (as they are in English), but rather distinct phonemes. In the same way, /k/ and /kh/ in กี่  /kìi/ and ขี่  /khìi/, /t/ and /th/ in ตั้ง  /tâŋ/ and ทั่ง  /thâŋ/, as well as /c/ and /ch/ in จาม  / caam/ and ชาม  / chaam / contrast on the basis of aspiration. Aspiration of voiceless initial stops and affricates is a distinctive feature of Thai phonology, but not of English phonology.

Likewise, vowel length shows itself to be a distinctive feature of Thai: ข้าว  /khâaw/ and เข้า  /khâw/, โคม  /khoom/ and คม  /khom/, and สีด /sìit/ and สิทธิ์ /sìt/.

It will be preferred here to use the older form of IPA transcription for Thai, in which ยืน  ‘to stand’ is transcribed as /jyyn/. Reverting to this earlier usage reduces by one—from five to four—the number of non-Roman characters needed to transcribe Thai.

The use of /j/ to denote the approximant, or semi-vowel, implied by the writing of ย is consistent with the historical derivation of ‘j’ from the ‘i’ of Latin. Accordingly, the character ‘j’ is used in the modern writing of Latin to indicate this very sound. It is also used to denote the Hebrew “yodh” (י) in the transliteration of Hebraic names that appear in the Hebrew Scriptures. Likewise, the orthography of German has also adopted the character ‘j’ for this sound.

/y/ in this system denotes the lower-high central-to-back vowel in ดึง / dyŋ / ‘to pull’.

Prosodic phonemes:

Less obvious, but important nonetheless, are distinctive contrasts in rhythm, stress and intonation contours. The prosodic phonemes of Thai are of these three types and will assume reality as we proceed.

Let’s suppose, now, that someone were to peer through the window and remark:

Two observations are readily apparent:

So, all the syllables uttered by the one peering through the window must be on the same pitch level, right? Well, not really. There is actually a gradual drop—a drop along a certain pitch line we call an intonation contour. The fact that the farmer’s behavior does not seem especially significant to the speaker puts his utterance along a “falling-intonation” contour.

Often, a speaker will vocalize his utterance along two or more such intonation contours, separated by what are called clause junctures. Let’s consider, first, a couple of English-language analogies.

1. “Close-clause juncture,” (English):
   
   “The weather in Phoenix during late June is unreal—it’s a test of human heat endurance.”
   
   Just after “…is unreal,” we pause briefly to start a new “breath group,” quite as if we just momentarily stopped talking and then started again at “it’s a test…”. The “breath group” is an intonation contour. After a close-clause juncture, we quickly start a new pitch contour, starting at the same pitch level we vocalized at the end of the previous contour. So, we start a new contour, saying “it’s a test…,” resuming at the same pitch at which we concluded the previous contour, saying “…is unreal.”


2. “Open-clause juncture,” (English):
   
   “It’s too hot here in late June. You really should visit Phoenix at some other time of the year.”
   
   After an open-clause juncture, we take the slight extra time to resume speaking at the same pitch level as we did at the beginning of the previous pitch contour. Both “It’s too…” and “You really…” start at the same level.

The entire sequence of syllables that are uttered along an intonation contour is called a “phonemic clause.”

Stress Phonemes:

Returning back to Thai, let’s consider:

The underscore /_ / indicates normal (loudness) stress. Normal stress is a benchmark stress level (a phoneme, and therefore meaningful) and is in contrast with weak stress (unmarked, no stress phoneme present) and loud stress (a phoneme indicated by a preceding /!/, albeit not applicable to this example).

Each of the underscored syllables above have normal diminishing stress, except /níi:/, which has normal sustained stress. Diminishing stress, whether weak, normal or loud, lessens in intensity from syllable onset to syllable final. Sustained stress is maintained at a generally constant (loudness) level throughout the phonation of the syllable.

In general, minor form-class members like prepositions, postpositions, conjunctions, most sentence particles and the more-common metric classifier will be found to have weak stress. Nouns, adjectives and verbs (other than those serialized after the main verb) will have normal stress in at least one syllable.

Reconsider the concept of the phonemic clause described above. A phonemic clause will typically have only one sustained stress /:/. After the sustained stress, all following syllables up to the juncture necessarily have only weak stress. In the sentence above, a close-clause juncture is indicated by the /,/.

It is actually possible to utter the exact same sentence along a single contour (corresponding to a single phonemic clause), with no break in the form of a juncture. Whether it is or not, the portion of a clause containing one sustained stress is called a “phonemic phrase.” After the sustained stress, all syllables have weak stress for the rest of the same phonemic phrase.

Rhythm Phonemes:

Let’s look at this same sentence again.

Notice that, in the expansion, adjacent syllables are separated either by a space, hyphen /-/, colon /:/, comma /,/ or no space or symbol at all. These indicate contrasts in rhythm, the relative time lapse between the onset of one syllable and the onset of the very next syllable. The speaker’s voice is not necessarily active throughout the entire time interval between the start of any two successive syllables. The time interval from the start of one syllable to the start of the very next is called the “duration” of that one syllable.

There are four rhythm phonemes which occur internal to a phonemic clause:

The two below are external:

The presence of either /,/ or /; / requires the existence of /:/ (sustained stress) on the last syllable just before the /,/ or /;/, unless some prior syllable is shown to have /:/ within the same phonemic phrase.

Quite appropriately, the last syllable uttered by the speaker is considered to have “infinite duration.” Its only significance is either that the speaker is waiting to hear what the listener has to say, or that their conversation is concluded.

As otherwise stated, rhythmic contrasts describe how fast we go from one syllable to the next. Additionally, just how fast we do go from one syllable to the next will change as we course through the sentence. The contrasts in rhythm correspond to these changes.

As we will now see, these contrasts in both stress and rhythm are often quite meaningful, as illustrated in the following contrasts: (The prosodic feature indicated by /^/ will be explained shortly.)

Notice that the occurrence of /,/ in the first of these requires ไหน  to have sustained stress: /nǎj:/, but with the longer duration of /,/.

On the other hand, consider:

The occurrence of the /,/ after /rə̌ə/ does not imply that /rə̌ə/ has sustained stress, since one of the preceding syllables in the same phonemic clause (namely, /níi:/) already has sustained stress.

The morpheme โรง  /rooŋ/ ‘facility’ actually has two variants (allomorphs): /rooŋ/ and /roŋ/. The short-vowel variant /roŋ/ occurs when its duration is medium-short /-/ and the preceding syllable has at least medium duration. Notice:

But:

(In the orthography, of course, this morpheme is always represented as โรง .) More generally, a long vowel morpheme will occur with a short vowel morph (or, allomorph) when its duration is medium-short and it is preceded by a syllable with medium or longer duration:

Further contrasts:

(Note that the first syllable of เวลา  / we-laa / is always short.)

(นา  is a variant of the sentence particle นะ  “Wouldn’t you say…” which occurs in the second relative position of the four-place clause-final codaphrase.)

We’ve seen two examples so far of a syllable with minimum (short) duration: /càk-krajaan /and /dâj kawe-laa /, in which /kra / (or /ka/) not only has short duration, but also weak stress. Such a syllable is also said to be “toneless”–it has no defined tone. While it is always true that a short-duration syllable will necessarily have only weak stress (and no tone), the converse is not quite true. Syllables with weak stress will often have more than short duration and a definitive phonemic tone.

Notice that เหรอ  / rə̌ə / has only weak stress, but long duration in

As we’ve noted, a syllable with minimum (short) duration will also be toneless. By “toneless,” we mean that a short-duration syllable does not have a specific pitch contour or level. Although the tone rules indicate that a syllable like กะ  or กระ  should have low tone, in a context like จักรยาน  /càk-krajaan / and ได้กับเวลา  /dâj kawe-laa/, in which it has minimum duration, it has no phonemic tone value. Note further that a toneless syllable (as actually pronounced) has no final consonant. Again, the converse is not true. Commonly, native speakers will phonetically apply a mid-tone in their pronunciations of such syllables.

Applying a low tone to กะ  or กระ  in จักรยาน  /càk-krajaan / or ได้กับเวลา  /dâj kawe-laa / will not impair understanding, provided that the speaker does not inadvertently add stress (louder than weak stress) or duration (longer than short duration) in his utterance of the syllable. However, it’s a lot easier to pronounce a minimum-duration syllable with a phonetic mid-tone.

Consequently, it can now be seen that the mid-tone is not some “default” tone occurring in the absence of some more definitive tone, but a particular tone phoneme in its own right, just like the low, falling, high and rising tones. A syllable like มี  /mii/ ‘to have, there to be,’ for instance, has a specific tone (mid-tone); it is not toneless.

Play the sound clip for the following adjective: อาคเนย์ . It should be noticed that the second syllable /คะ/ is not sounded (phonetically) with a high tone (as the tone rules would indicate) but (phonetically) at a mid-tone level. Phonemically, this syllable is toneless: อาคเนย์  /aa-khanee/, in which คะ  /kha / has short (minimum) duration and, accordingly, weak stress.

It was parenthetically noted, just above, that the first syllable of เวลา  ‘time’ is characteristically short, though written as if long. This first syllable is not, of course, a constituent morph with any separate meaning of its own,

Intonation Phonemes:

Let’s suppose, now, that the same person who peered through the window before were suddenly to turn to those inside and say excitedly:

Each of the syllables is still mid-tone as before. The intonation contour, however, is not falling as it was when the farmer was merely watching the fish, but maintains the same level as it had at the first syllable. The level is not only constant, but slightly higher than it would be at the beginning of a falling-intonation contour. The action of the farmer is regarded as remarkable, significant or bizarre by the speaker and causes him to utter his observation along a constantly level, slightly elevated, intonation contour. Further, the net volume of his utterance is somewhat greater than it was when he noticed that the farmer was only watching the fish. However, the relative loudness (stress) levels of medium, weak and loud (if any) are still applicable:

The /^/ is a phoneme which indicates that the pitch line is being “held up” by the speaker all the way to the end. It is also the sole constituent of a morpheme conveying momentary excitement or an impression at how remarkable, significant, notable and/or bizarre is the situation. It is added by the speaker to say that he is suddenly impressed or affected with a certain sense of excitement in some particular way. In the absence of this phoneme, the contour is (by default) falling (not a phoneme).

The particles หรอก  and น่ะ  in the example above fill the first and second relative positions of a four-place codaphrase that commonly terminates a Thai sentence clauses. หรอก  at the end of negative questions expresses the attitude ‘how could it not be true?!’ and น่ะ  the attitude ‘ well, don’t you think so?’ Also, /khãw/ has the plain high tone, as will be discussed farther down.

Now, let’s reexamine that twin-clause example we saw before:

There are two falling contours in this example of close-clause juncture.

If we replace the /,/ with /;/, we get:

The two clauses are now in open-clause juncture. The duration of the weakly-stressed เหรอ  / rə̌ə /is extra-long. Notice that [the two clauses] X [the two pitch contours] X [the two clause junctures] = [eight intonation sequences]:

1) Two falling contours in close-clause juncture:

2) Two falling contours in open-clause juncture:

3) High-level and falling contours in close-clause juncture:

4) High-level and falling contours in open-clause juncture:

5) Falling and high-level contours in close-clause juncture (relatively infrequent):

More likely, in this instance, would be a single, high-level contour:

6) Falling and high-level contours in open-clause juncture:

7) Two high-level contours in close-clause juncture:

8) Two high-level contours in open-clause juncture:

When the first clause is uttered along a (slightly elevated) level contour with extra net loudness, the speaker finds it remarkable that “this [could actually] be the bike” (that concerns the listener). When the second clause contains this contour, the speaker finds it remarkable that the listener actually “rode it to school.”

Concept of Morpheme and Lexeme:

We review, first, the concept of the morpheme: the smallest meaningful unit of speech. จักรยาน  / càk-krajaan / ‘bicycle’ comprises the two constituent morphemes: Of the first, จักร  /càk / ‘wheel, disc, cycle’ ➔ / càk-kra / as first of a compound, ➔ /càk / elsewhere. The morph / càk-kra / is said to be “bound” to its environment within the compound and cannot stand alone. The second is ยาน  /jaan/ ’vehicle.’ Consider also เครื่องจักร /khrŷaŋ-càk / ‘machine(ry), engine’ and ยานอวกาศ / jaan à-wakàat / ‘spacecraft’.

Morpheme /càk / has the two allomorphs shown, of which one is bound and the other free. Accordingly, we say that the “constituent morphs” of /càk-krajaan / are /càk-kra / and / jaan /. /càk-krajaan /, we say, is a “lexeme,” a complete “word” in its own right.

Notice, too, that the English word ‘bicycle’ (a lexeme) consists of two constituent morphs ‘bi-’ and ‘cycle.’ The first, ‘bi-’, is a bound morph, as it cannot occur independently from some other morph, like ‘cycle.’ The second, ‘cycle,’ is an independent morph; it needn’t attach itself to some other morph, as does ‘bi-’. We recognize ‘bicycle’ as a lexeme by the fact that, when ‘bicycle’ is used in a sentence, its constituent parts are never separated when ‘bicycle’ is repositioned to a different part of the sentence:

Notice that it is not possible to reposition just the ‘bi-’ and not also the ‘cycle’, nor to reposition just ‘cycle’ and not also ‘bi-’.

We also recognize ‘bicycle’ as a lexeme in that, if a new word (like ‘red’, for instance) is inserted into the sentence containing ‘bicycle’, the new entry can never be inserted between its two constituent morphs. No matter how many constituent morphs a lexeme contains, insertion of a new morph or lexeme in the sentence can never result in separation of any two (or more) of the meaningful parts of a given lexeme.

The English lexeme ‘reconstruction’ contains four constituent morphs (‘do over’ + ‘together with’ + ‘assemble’ + ‘express as noun’), all of which are bound. If any one of these is relocated within a sentence, they are all relocated and maintain the same construction. If anything is inserted into the sentence, it never interposes between any two of these morphs.

Prosodic Morphemes:

Morphemes (meaningful units) in Thai are composed, not only of mixes of consonants and vowels, but also of prosodic units consisting of stress, rhythm and intonation contours.

The morpheme /;/, meaning ‘at the end of one sentence clause, ready to start another’ contains only the open-clause juncture phoneme (;).

The sentence above contains two phonemic clauses, but only one syntactic clause. Inserting /;/ changes both meaning and structure:

This latter sentence contains two phonemic clauses and two syntactic clauses.

Note that the syllable / wít / in /khɔ̌ɔŋ khun-suwít: , khãw fàak wáj/

and in /khɔ̌ɔŋ khun-suwít: ; khãw fàak wáj/ has normal sustained stress since no preceding syllable has sustained stress. Its duration, however, is long /,/ and extra-long /;/, respectively, overriding the medium-long duration associated with sustained stress. Compare:

Here, เหรอ / rə̌ə / has weak diminishing stress and long duration, while นี้ /níi:/ has normal sustained stress with medium-long duration.

Meaningful Contrasts of Stress, Rhythm and Pitch:

ชั่ง / châŋ / is a (weak-stressed) modal meaning “really loves to do, can’t resist doing.”

(Note the effect of the presence of the high-level contour ^.)

ช่าง  / châaŋ / ‘artisan, craftsman, technician, specialist’ often enters into compounds like ผมเป็นช่างไม้ /phõm pen châŋ-máaj/ ‘I’m a carpenter’ (ไม้  of ช่างไม้ is long.)

Recall that a long-vowel morpheme (ช่าง  /châaŋ /, for instance) will occur with a short-vowel morph (or, allomorph) when its duration is medium-short and it is preceded by a syllable with medium or longer duration.)

Review example of ช่างผม /châaŋ-phǒm / ‘hairdresser/stylist’ here.

ช่างพูด has at least these two manifestations:

We see from the foregoing that stress and rhythm are not determined. Generally, in the context of a given series of syllables that make up a valid Thai expression, stress and rhythm cannot be predicted. They may be changed in such a way as to change the meaning, either to a different meaning or to no meaning. Likewise, the syllable tone can be changed independently, resulting in either a new meaning or no meaning at all. Accordingly, we say that stress and rhythm are distinctive features of Thai phonology, as is the tone.

Notice, too, that morpheme ช่าง ➔ / châŋ / when duration is medium-short and preceding syllable has at least medium duration, but ➔ / châaŋ / otherwise. Thus, ช่างไม้ ➔ / châŋ-máaj /, but ผมเป็นช่างไม้ ➔ /phõm pen châŋ-máaj /. Again, compare with the example of ช่างผม ‘hairdresser’ here.

As an aside, consider the well-known expression of thanks, ขอบคุณ /khɔ̀p khun/. Corresponding to this familiar response, we typically say “thank you.” However, do the individual parts as expressed in each language correspond to one another? We should realize that the stressed /khun/ does not here mean ‘you,’ but ‘favor, kindness,” while the verb /khɔ̀p/ means ‘to be grateful, to repay.” The literal meaning of ขอบคุณ is thus “to acknowledge the kindness or favor.” ขอบคุณ may, however, be followed by a weak-stressed second-person pronoun indicating the person being thanked.

We saw under the heading Prosodic Morphemes clause-juncture contrasts of /,/ and /;/ separating the two phonemic clauses:

ของคุณสุวิทย์ /khɔ̌ɔŋ khun-suwít: / and / เขาฝากไว้ /khãw fàak wáj/

Both of these clauses were uttered on falling-pitch contours.

Now consider:

Here, both clauses are uttered on high (level) contours, indicating a sense of excitement on the part of the speaker. Also:

Notice that, in the first phonemic clause / !mii /has loud diminishing stress and /ry̌y: / has weak sustained stress. This first clause is uttered on a high contour /^/ ‘momentary excitement.’ The second is uttered on the more normal falling contour. Compare with:

We see a similar pattern, but with close-clause juncture, in:

นิ่ / nî / and ค่ะ /khâ /are weakly stressed sentence particles that occur in the first and third relative positions, respectively, of a clause-final codaphrase. นิ่ means: “Take note of this matter.” ค่ะ is the statement form of a female sex-marker particle that refers back to the speaker.

With both clauses on high contours in open-clause juncture, the above becomes:

Here, we have two clauses uttered on high-level pitch contours that are in open-clause juncture.

These examples remind us that the morpheme /; / (open-clause juncture) means ‘at the end of one sentence clause, ready to start another’ and consists only of the one phoneme /; /.

We notice that the semantic content of the morpheme consisting of only /, / ‘at end of pitch contour, ready to start another’ is less than that of /; /. Still, the insertion of /, / where there otherwise is no clause juncture at all necessarily does affect sentence structure and ultimately does reshape the meaning. Consider for instance:

(First and second-person pronouns in conversational Thai are usually replaced by actual names or, more frequently, nicknames. Unlike the pronouns, the names and nicknames are stressed. The resulting sentences appear to have third-person subjects.)

ล่ะ  /lâ/ is a sentence particle meaning ‘I want the answer.’ It occurs in the second relative position of the four-place codaphrase terminating some Thai sentences.

The meaning contrast is more significant here:

มา  / maa / here is a verbal postposition to the main verb ไป  /paj /and indicates that the action of the main verb is being extended from the past.

มา  /maa / here is the name of the listener. It is the sole constituent of the vocative phrase, uttered on a separate pitch contour of its own, following the close-clause juncture /, /.

Meanings of short responses to simple questions can be influenced by the presence or non-presence of /^/ ‘momentary excitement’:

Syllables normally short become long under conditions of loud sustained stress. Consequently, the morpheme /caj/ has at least these two allomorphs:

/caj/ ➔ [caaj] in presence of loud sustained stress, ➔ [caj] elsewhere.

Alternatively:

หรอก  / rɔ̀k/ is a sentence particle meaning’ definitely not.’ It occurs in the first relative position of the four-place codaphrase terminating some Thai sentences.

Rhythm and Stress Morphemes:

Thus we have a morph consisting only of the normal-stress phoneme. This morph is actually an allomorph of the cardinal numeral หนึ่ง /nỳŋ / (or /nyŋ /) ‘one’.

Consider, for instance,

Here, the classifier / khon / has its typical weak stress. / ìik-sɔ̌ɔŋ / is a prepositional phrase meaning ‘two more.’ The same is true of the following phrase:

Both / sɔ̌ɔŋ / and / nỳŋ / in these constructions are numerals.

Yet another way to say ‘One more member’ is:

Notice that, by adding the morpheme /_/ ‘a single instance’ to the (otherwise unstressed) classifier /khon/, we get: /khon/ ‘an instance [of /samaa-chík/]’

In the last two examples, the intended meaning is specifically ‘one more member,’ and not ‘two’ or ‘three.’ If the intention is simply ‘another member’ and we don’t mean precisely ‘one’ (as opposed to ‘two’ or ‘three’ members), then instead of the numeral /nỳŋ/, the unstressed demonstrative /nỳŋ/ is used:

/nỳŋ / here is a member of a subset of demonstratives with class meaning ‘specifying or generalizing among all possibilities.’ The other ‘better-known’ subclass, having familiar members like / níi /, / nán / and / nóon / (and a few others), means ‘specifying by relative location’— relatively close to, or distant from, the speaker.

The unstressed / khon /, with or without the added morph /_ / ‘a single instance,’ is a unit classifier—a classifier whose antecedent is considered to a discreet unit and thereby indivisible. Nouns that denote discreet units are called “count nouns.’ Unit classifiers are semantically tied to corresponding classes of discreet items or cont nouns (as does / khon / to nouns denoting people).

metric classifiers have almost no such semantic relationship. They denote amounts, sets and values and, as such, often occur in construction with nouns that are thought to be divisible, or “mass nouns,’ metric classifiers are usually unstressed counterparts (termed “allolexes”) of nouns that commonly carry normal stress.

The normally-stressed noun / thûaj / occurs here. The normal stress is simply part of its phonological make-up; it is not a separate phoneme.

The unstressed metric classifier / thûa j / occurs in this classifier phrase.

Here, the unstressed metric classifier / kɛ̂ɛw / occurs together with the morpheme /_ / to indicate a single instance of this same classifier. Note that the metric classifier with stress added is uttered on a new pitch contour. There are two falling-pitch contours in this last example, separated by close-clause juncture /,/.

Other examples:

Here, / sɔɔŋ / ‘pack, envelope’ occurs as a noun with normal stress.

The first occurrence of /sɔɔŋ / is the metric classifier for items contained in packs. The second of / sɔɔŋ: / is the regularly stressed noun ‘pack.’ / muan / is the unit classifier for cigarettes. /sɔɔŋ: lã / is a postpositional phrase containing the postposition / lã / ‘per.’ Note that there are two clause junctures: the first open /;/, the second close /, /.

As a morpheme, /!/ is an imperative or at least an exhortation to action.

The verb กิน here is uttered as a command.

In transcription, กิน  ! (as a command) would appear as /! kin/. The initial onset of /k/➔/ i / is noticeably louder than normal. The loudness quickly subsides throughout the phonation of /i/ and, as /i/ ➔ /n/, the syllable actually becomes voiceless, so that the /n/ is mostly or completely uttered as when whispering. Likewise,

In transcription, ไป  ! as a command is rendered as /! paj/, in which the sequence of /p/ ➔ /a/ ➔ /j/ involves strong initial loudness /p/ ➔ /a/ rapidly diminishing to (partial) voicelessness /a/ ➔ /j/ on the semivowel (or approximant) /j/.

/ ! :/ ‘Most certainly does! ; does too!’ or ‘Of course it does! They sure do!’

When / ! : /is present, speaker expresses a contradictory assertion of his own or an insistence on the probable truth of his assertion:

Recall that loud sustained stress on a syllable requires that it be long. Thus,

/kin/ ➔ [kiin] in presence of loud sustained stress, ➔ [kin] elsewhere.

/ !… :/ expresses a contrastive emphasis among possible or actual alternatives. It occurs with a sequence of one or more syllables. In the special case of its co-occurrence with only one syllable, it is a homonym of / ! :/ ‘Does too; of course it does!’ / !… :/ is often followed by the morpheme /^/ ‘momentary excitement,’ consisting only of the prosodic phoneme /^/ (high-pitch contour).

Notice that /!diaw: / has loud sustained stress, /!sǎam / has loud diminishing stress and /lêm:/ has weak sustained stress. The morpheme / !… :/ co-occurs with the syllable sequence สาม เล่ม .

Recall this example:

This morpheme /: /consists only of the sustained-stress phoneme /:/.

Notice that the second (and last) phonemic clause /ciŋ-côo cajyyn-khajèŋ dâj / contains no /: /. The last phonemic clause of a sentence may or may not contain a /: / if its pitch contour is falling.

We’ve already seen the following contrast:

The sustained stress /:/ here indicates last stressed syllable (meaning syllable with at least normal stress) till end of utterance. If a syllable with sustained stress is not immediately followed by either a close or open clause juncture, or end of utterance, then the duration of the syllable is medium-long, as in / paj nǎj: maa /. When followed by /,/, /nǎj: / has long duration:

/:/ indicates there will be no further stressed syllables (or that all remaining sylabes are weak) till next /, /. Since /nǎj: / here is followed immediately by close clause juncture /, /, the duration of the syllable is long.

The morpheme /- / consists only of the medium-short rhythm phoneme /-/. (Review section “Concept of Morpheme and Lexeme,” at bottom of page 12 above.)

/-/ can also mean ‘linking two adjacent minor form-class lexemes into a larger collective thought.’ Compare:

(เสีย / sǐa / here is a verb, a major form-class lexeme.) with:

(เสีย  /sǐa / here is postposition to the verb ทำ  / tham, / and serves to define its aspect as ‘action seen as a completed unit.’ It is often pronounced as though it were written as ซะ  /sã /.)

It cannot be assumed that all occurrences of /-/ indicates the morpheme just defined. /- / is sometimes only a constituent phoneme of a lexical morph. Consider:

เก้าอี้  / kâw- îi / ‘chair,’ of which /- / is simply part of its phonological make-up, but is not a meaningful element in its own right. The two syllables that make up /kâw- îi / are not meaningful constituents (or parts) of some larger word-like entity meaning ‘chair,’ joined to each other by /- /. As otherwise stated, the lexeme เก้าอี้  /kâw-îi / is not further analyzable (cannot be further subdivided) into smaller meaningful parts (or morphs). /- /, a phoneme, merely indicates that the syllable /kâw-/ in / kâw-îi / has medium-short duration.

The two lexemes ต้นสัก / tôn-sàk / and ตั้งต้น / tâŋ-tôn /, do, by contrast, consist of three constituent morphs each, two of which are syllabic and the other prosodic. The syllabic morphs / tôn /, /sàk / and / tâŋ / are each lexemes (or, “words”) in their own right, consisting of only one constituent morph each. They are thus called “free morphemes.” The significance of the prosodic morph /-/ is that it is joining two morphemes that could otherwise stand as independent lexemes. No such joining occurs in เก้าอี้  / kâw-îi /.

Occasionally, a free morpheme will have an allomorph that is not free, but occurs only in combination with some other morph, which may or not itself be free. An example is วิทยา  / wít-thajaa / ‘knowledge, science,’ which has an allomorph /wít-thaj… /. This morph consists of two syllables (/ wít-tha /) plus a fraction of a syllable (/ j… /). Obviously, it cannot stand alone. It combines very well, however, with อาลัย  /aa-laj / ‘dwelling, residence, abode’ to form the lexeme วิทยาลัย  / wít-thajaa-laj / ‘college’. Note here that the second syllable, which has only minimal (or short) duration, is toneless. Note also that the long vowel /aa / seems to “belong to”, or to be “shared by”, both / wít-thajaa / and / aa-laj /.

We just saw วิทยาลัย  / wít-thajaa-laj / ‘college.’ Now, how about มหาวิทยาลัย

/máhǎa wít-thajaa-laj / ‘university’? The / / places / máhǎa / ‘great, large’ and /wít-thajaa-laj /on the same hierarchical level and joins them together into a still larger unit. On a yet-lower level, /-/ links / wít-thaj… / and / aa-laj / to make /wít-thajaa-laj /.

We also saw:

In both if these examples, / / links two higher-order constituents, one of which subdivides into two lower-order meaningful elements (morphs) joined by /-/ into a lexeme. In the first, / / links verb with object and, in the second, auxiliary verb with main verb. / / is meaningful in that it aids us in distinguishing these higher-ordered constituents from each other. Further, the contrast with /-/ aids our recognition of those that are joined at a lower order into one of these higher-ordered units.

Contrasts of / / and /-/ are especially helpful in defining compound numeral lexemes consisting of multiplications and additions of simple numerals:

แปดร้อยเจ็ดสิบหก / pɛ̀ɛt-rɔ́ɔj cèt-sìp hòk / ‘876’

Here we see that /-/ joins two lower-ordered morphs for the purpose of multiplications while / / joins two higher-ordered lexical units for the purpose of addition.

As regards the significance of / /, consider the following:

As before, the contrast of / / with /-/ helps to subgroup the reduplicated instances of / dèk / into a larger lexeme and to distinguish this larger lexeme from /cakin /.

From the example above, it’s clear that we could also have

We thus wish to modify the definition of / / to say:

Sentence patterns like แม่จะกิน / mɛ̂ɛ cakin / motivate this modification. Although จะ  / ca / is written to imply a low tone, it occurs only with weak stress in tight constructions with verbs and adjectives. Having only minimum (or short) duration, จะ  /ca/ is necessarily toneless.

       
When we combine the morpheme / / above with /-/ and with /:/, we get some interesting contrasts:

(กัน / kan / is a verb, เหลือ / ly̌a / is a name.)

(กัน /kan / is the reciprocal pronoun.)

(ช่วยเหลือ / chûaj-ly̌a / is all one verb.)

Of course, there is no contrast in the orthography. Neither would there be any contrast in the transcription, except for the inclusion of the prosodic features.

The Sixth Tone:

Additionally, notice the occurrence of the plain high tone on the pronoun เขา  / khãw / ‘he, she, they.’ Weakly stressed pronouns ผม  / phõm / ‘I (masc.) and ฉัน  / chãn / ‘I (fem.) and also one allomorph of the preposition ที่  / thîi /, namely / thĩ /, ‘to, at’ contain the plain high tone as part of their phonological make-up. The plain high tone belies the rising and falling tones implied in the orthography of ผม , ฉัน , เขา  and ที่ .

The plain high tone is a slightly elevated mid tone, occurring with no noticeable vocal constriction. The constricted high tone, which is much more common, begins relatively quite high and increases linearly in frequency to outright vocal constriction throughout the duration of the syllable.

In prepositional phrases, however, the pronouns indicated above occur in stressed rising-tone variants, or “allolexes,” consistent with their orthographies. A variant of a lexeme is called an allolex, where a lexeme is an independent word. Thus:

Notice that these rising-tone allolexes of the pronouns also have normal stress. Notice, too, the proposition ของ  / khɔ̃ŋ / ‘of, belonging to’ has weak stress and that its tone is the plain high tone. Notice, too, that in รถของผม / rót khɔ̃ŋ-phǒm / ‘my car,’ the morpheme / / described above defines the lexical hierarchies of / rót / and / khɔ̃ŋ-phǒm / as equivalent.

It should also be noted that pronouns ผม , ฉัน  and เขา  will occur in yet another allolex when the preposition indicating possession is omitted. In this case, the indicated rising tone is accompanyed by weak stress:

On the other hand:

Likewise, the first syllable of หนังสือ  / nãŋ-sy̌y / ‘book(s), papers, reading and writing’ will often contain the plain high tone (rather than the indicated rising tone). Nonetheless, หนังสือ  is also rendered as the allolex / nǎŋ-sy̌y /. / nãŋ-sy̌y / and / nǎŋ-sy̌y / are in free variation in that the speaker might choose either variant practically at will, without regard to the immediate environment of its usage. By contrast, compare the allolexes of the pronouns considered just above.

The allolexes / thîi / and / thĩ / ‘to, at’ are also in free variation:

Some comments are due this last example. There are two pitch contours, both falling, in close-clause juncture.

/ chǎn : / is required to have sustained stress before the /, / since no previous syllable along the same pitch contour has sustained stress.

Enumerations consist of numerals (like / hâa-rɔ́ɔj /) plus a classifier (like the metric classifier / bàat /). In classifier phrases, the classifier has weak stress.

Additionally, the antecedent of classifier / bàat / is เงิน  / ŋən / ‘money, silver’, which does not appear explicitly. Its presence is implied as if the sentence had actually been:

We already saw this example of a classifier phrases:

Notice here that the postposition / lã / ‘per,’ as well as the pronoun / phõm /, has the plain high tone.

A number of clause-final sentence particles have the plain high tone. In general, these particles are arranged in an ordered sequence of one to four (mostly) weakly-stressed lexemes expressing speaker’s attitude toward the subject and context in which he is speaking. This sequence is called the “codaphrase.”

Among the particles that can occur in the first relative position of the codaphrase, is ละ  / lá /, / l ã / and / la /. Its meaning is: ‘new course of action.’ /lá/ occurs when not followed by other particles; / lã / or /la / occurs otherwise.

The particle / rə̌ə / occurs in second position, asking objectively: “true or not?”

/ nã / occurs in the second relative position with meaning ‘all right, now?’; ‘it would seem so’; ‘wouldn’t you say?’ This particle is generally observational and occurs in a number of variants, the most common of which is with the plain high tone and weak stress.

The particle / mãj / occurs in first relative position, asking subjectively: “do you find it to be true?’ Whether written as ไหม or มั้ย , it is almost always uttered with the plain high tone as shown. Only in pre-scripted broadcasts and in certain formal functions is it ever uttered as / mǎj /.

รึ  / rỹ /, like the particle / rə̌ə /, is a variant of หรือ  / ry̌y /, objectively asking: ‘true or not?’ หรือ  /ry̌y/ is used in the same formal contexts as ไหม / mǎj / above.

Notice, too, that the pronoun / khãw / is distinguished from / khǎw / ‘hill, mountain’ by the tonal differences of plain high and constricted high tone. The phonemic status of the plain high tone is thus established by the minimal pair of /khãw / ‘he, she, they’ and / khǎw / ‘hill, mountain’. (Pronoun / khãw / is written more formally as เขา , but informally as either เค้า  or เค๊า.)

Here, there is no particle in the first relative position of the codaphrase.

ละ  / lã / occurs as a variant of ล่ะ  / lâ / in second position with the meaning ‘I must know!’ The variant / lã / is chosen when followed by other sentence particles.

/ khã / is the question form of the female sex-marker particle and occurs in the third relative position of the codaphrase. It refers back to the speaker. (The statement form is ค่ะ  / khâ /.)

In these two examples, ครับ  /khrãp / (with plain high tone) is the question form of the male sex-marker particle, and / khráp / (with constricted high tone) is the statement form. Like / khâ / and / khã /, they occur in third position. In the orthography, /khrãp / and / khráp / look alike (ครับ ), though they differ by tone.

The inclusion of the second-position particle นะ  / nã / in the first example results in the English as it appears. In the absence of the particle, the English would be: ‘Which way are you going?’

Additional Contrasts:

We conclude this paper by examining some additional meaningful contrasts brought about by the inclusion of prosodic morphemes, only. These contrasts will be presented in three groups and, with the exception of the orthography, are extracted from the Reference Thai Grammar of Richard B. Noss (1964). The first or these contrasts the usage of the two rhythmic morphemes:

/- / ‘joining two morphs together within a major form-class lexeme; linking two adjacent minor form-class lexemes into a larger collective thought’

and

/ / ‘joining two upper-level (or higher-order) units, at least one of which contains lower-order units (normally morphs) joined by /-/ or simply joined with minimum (short duration).

(The definitions of these two morphemes, as well as most others in this paper, are not taken from Noss.)

Note that, while the two examples above differ in their orthographies, they would otherwise sound alike were it not for contrasts of rhythm and stress.

The second group contrasts the usage of normal and weak stress:

/ klàp / in the second of this pair is a modal meaning ‘acting unexpectedly or reversing previous behavior’

Finally, the third group contrasts the usage of the phonemic clause boundary /,/:

/ lá / is the same sentence particle described here.

Phonetically-Only Contrasts:

The plain high tone described above is a phoneme because its actual occurrence is not determined, as demonstrated by the existence of minimal pairs that differ only by the presence or non-presence of this tone. Thus, unlike the two topics discussed below, the contrast presneted by the plain high tone is not merely phonetic, but genuinely phonemic.

Relevance of the Glottal Stop:

A certain phonetic segment called the glottal stop [ʔ] is not, however, a phoneme, at least not in Thai. It is generated by the very momentary closing of the vocal chords and a simultaneous rapid increase in lung air pressure bearing against this closure. The air then bursts out, creating a noticeably abrupt sound.

In colloquial English, we use the glottal stop to distinguish the affirmative response “uh-huh” /ʌ-hʌ/ ‘that’s right’ from the negative “uh-uh” /ʔʌ-ʔʌ/ ‘no way!’ Thus, the /ʔ/ is indeed a phoneme in our spoken (US) English.

In Thai, the occurrence of [ʔ] is determined by its immediate environment. In the absence of a any other initial consonant, a stressed syllable (having normal or loud stress) will be uttered with a leading [ʔ]:

In วิ่งออก /wîŋ ɔ̀ɔk/ ‘ran outside’, the verbal postposition / ɔ̀ɔk / has weak stress and is not preceded by[ʔ]. Phonetically it is [ɔ̀ɔk], symbolically no different from its phonemic transcription of / ɔ̀ɔk /.

In อ่านออก /àan ɔ̀ɔk / ‘able to read it’, however, the completive (or resultative) verb / ɔ̀ɔk / has normal stress and is uttered with a leading [ʔ]. Accordingly, its phonetic transcription is [ʔɔ̀ɔk].

[ʔ] occurs in syllable-final position only if the vowel is short and the syllable is stressed. Thus:

โต  / too / ‘grown; big, large’ does not terminate with a glottal stop, while โต๊ะ  / tó / ‘table’ does. Phonetically,โต๊ะ  is transcribed as [tóʔ]. On the other hand, in หล่อนสวยนะ  / lɔ̀n sǔaj: nã / ‘Pretty, isn’t she?’, the observational particle / nã /, although short, has only weak stress. It does not, for that reason, terminate with [ʔ].

Lenis-Fortis Allophones:

The pair จำ  /cam / ‘to remember, recognize; to bind, compel’ and จาม  /caam/ ‘to sneeze’ seem to vary only by their contrasting vowel lengths. Vowel length is thus shown to be a distinctive feature of Thai phonology. /cam/ and /caam/ do indeed constitute a minimal pair. Phonemically, they differ only by vowel length.

Phonetically, however, there is a difference in the particular way in which each of these two verbs is actually uttered. The difference occurs automatically as a result of the difference in vowel length and affects the emphasis placed on the final consonant /m /. In /cam /, the speaker unconsciously applies an extra-firm, somewhat prolonged, closure of the lips to the creation of the (bilabial) nasal continuant /m /. The /m / pronounced as such is said to be ‘fortis.’

Short, stressed syllables with finals—especially the nasal continuants /m/, /n/ and /ŋ/—involve this more marked pronunciation of their finals. (Those familiar with Italian may have noticed that all syllable-final consonants are fortis.)

The final /m/ in /caam / is said to be “lenis” and, as such, is uttered with relatively weak muscular tension of the two lips. Finals terminating long vowel syllables are lenis.

If we denote the allophones of /m / as [m] (lenis) and [m̩] (fortis), then we would say: /m/ ➔ [m̩] as final of short, stressed syllable; ➔ [m] otherwise.

Note that the [ʔ] discussed above is the fortis glottal stop. Its occurrences are not relevant to the allophones of consonant phonemes occurring in final position just described.

This article created August 2012