Languages of the world differ in the number of different morphemes that are used, on average, in building words in the language. While MSEA languages are often said to provide the paradigm cases of isolating morphological profile, it is important to understand what this means. If a language were to show the logical extreme of the isolating profile – with a morpheme-to-word ratio of exactly 1:1 – the language would have no word-formation processes to speak of. MSEA languages can be regarded as isolating only in so far as they are further toward the isolating end of the spectrum than most languages of the world. This does not mean that they have no morphology. As we shall see in this chapter, MSEA languages have ample morphological resources for forming words.
Half a century ago, Kratochvíl (Reference Kratochvíl1968) debunked the myth that all words in standard Chinese are monomorphemic monosyllables. He offers a five-way distinction in ‘basic types of morphological construction’, shown in Table 4.1.
Table 4.1 Kratochvíl’s five-way distinction of ‘basic types of morphological construction’ in Mandarin Chinese (SN)
| Word type | Token frequency rank, relative to other types | Example |
|---|
| Simple monomorphemic monosyllabic | 1 | huǒ ‘fire’ |
| Compounds (polymorphemic polysyllabic) | 2 | huǒ-chē ‘train’ (fire-vehicle) |
| Monomorphemic polysyllabic | 3 | húdié ‘butterfly’ |
| Polymorphemic polysyllabic | 4 | ré-du ‘temperature’ |
| | (hot-DEGREE) |
| Polymorphemic monosyllabic | 5 | maō-r ‘cat-DIM’ |
Compounds are the second most frequent of these word types in texts in conversation (Kratochvíl Reference Kratochvíl1968: 64). The category of ‘compound’ belies a diversity of underlying structures. Even in his ‘elementary outline’ of compounds in Chinese, Kratochvíl (Reference Kratochvíl1968: 73–79) devotes seven pages to explicating the possible structures, under four headings: coordinate compounds, subordinate compounds, reduplicated compounds, and stump compounds. This barely scratches the surface of the depth and breadth of compounding in Chinese.
The case of Mandarin is offered here in order to illustrate the degree to which a supposedly isolating language can in fact show a diverse range of ways in which words can consist of multiple morphemes. Given the profile of MSEA languages more broadly, we should expect the same situation to pertain in similarly ‘isolating’ languages of MSEA. The take-home message here is that while the languages of MSEA can be placed toward the isolating end of a continuum, it is a myth that the languages show the logical extreme of an idealized isolating/analytic typological profile.
4.1 Form Classes
In order to build words, phrases, and sentences, a speaker of any language must start with a set of form classes. The most important of these are the familiar word classes such as noun, verb, adjective, and preposition. There are many other types of word class in addition, for example ideophones, discourse particles, and numeral classifiers (see below sections of this chapter).
We can first make some generalizations about the major types of form class found in MSEA languages.
4.1.1 Noun–Verb Distinction Is Usually Clear
MSEA languages typically show a clear distinction between the noun and verb categories. There is of course some flexibility, with certain words occurring as both nouns and verbs, with related meanings. Consider, for example, the Lao word mok2. As a noun, mok2 is ‘a dish made in a package by steaming’, while as a verb it means ‘to prepare a dish by steaming ingredients in a package’. This word belongs to a set of words in Lao that show regular alternation, in a similar way to the set of English words that can be used either as nouns denoting tools or as verbs for denoting actions done by those tools: e.g., spoon, shovel, hammer. But in the main, in MSEA languages, nouns are used as nouns and verbs are used as verbs, with clear differences in grammatical behaviour in the verbal and nominal domain. Members of a verb class will take an array of markings, including aspectual-modal distinctions such as perfect, experiential, and negation. Members of a noun class will show other grammatical behaviours, including their use in numeral classifier constructions. See Chapters 5 and 6 for overviews of the nominal and verbal domains in MSEA languages.
4.1.2 ‘Adjectives’ Are Verbs, or Are Verb-Like
In some languages, an adjective class will be entirely distinct from both the noun and verb classes. But in many languages the adjective class will either be noun-like or verb-like, meaning that adjectives will resemble either nouns or verbs in terms of their grammatical behaviours. In English, adjectives happen to be distinct in their grammatical behaviour from both nouns and verbs. In a noun-like adjective language (such as Latin), the adjective ‘old’ might be rendered noun-like as, roughly, ‘an old one’, while in a verb-like adjective language, such as Sedang (Smith Reference Smith1979: 84), or any other MSEA language, it might be rendered verb-like as, roughly, ‘to be old’. Adjectives in MSEA languages are clear examples of the verb-like type, meaning that they show many of the same basic grammatical behaviours that verbs do (Comrie Reference Comrie and Kullavanijaya2007: 41).
Post (Reference Post2008) investigates the status of the adjective concept in Thai. He highlights the analyst’s role in determining word-class categories, through making useful generalizations about the data rather than discovering an objectively privileged taxonomic organization in the language. His data show that ‘adjectives’ may be understood not as a separate word class in Thai, but as a semantically defined subset of verbs (‘property concept’ words; Dixon Reference Dixon2010). In Thai, a verb, such as dəən ‘walk’, acts as a predicate without any marking (Post Reference Post2008: 343):
(4.1)| khon | níi | dəən |
| clf:person | prx | walk |
| ‘This person walks.’ (active verb, intransitive predicate) |
| Thai | TK | Post Reference Post2008: 343 |
By contrast, if a noun, such as phráʔ ‘monk’, is to be used as a predication, it must appear as the complement of a copula verb (Post Reference Post2008: 343):
(4.2)| khon | níi | pen | phráʔ |
| clf:person | prx | acop | monk |
| ‘This person is a monk.’ (concrete noun, copular complement) |
| Thai | TK | Post Reference Post2008: 343 |
Now consider the behaviour of the property concept word dii ‘good’. The following examples show that dii ‘good’ follows the same pattern of grammatical behaviour as the verb dəən ‘walk’, and not that of the noun phráʔ ‘monk’:
(4.3)| khon | níi | dii |
| clf:person | prx | good |
| ‘This person is good.’ (verb-like property term, intransitive predicate) |
| Thai | TK | Post Reference Post2008: 343 |
Another grammatical property – the use of the relativizer thîi – similarly distinguishes nouns from verbs, where dəən ‘walk’ is joined to a nominal head using the relativizer, while a noun must, again, be hosted by a verb, here the copula pen:
(4.5)| khon | thîi | dəən |
| clf:person | rel | walk |
| ‘person who walks’ (active verb, intransitive predicate) |
| Thai | TK | Post Reference Post2008: 344 |
(4.7)| khon | thîi | pen | phráʔ |
| clf:person | rel | acop | monk |
| ‘person who is a monk’ (concrete noun, copular complement) |
| Thai | TK | Post Reference Post2008: 344 |
Again, the property concept word dii ‘good’ shows the same pattern of grammatical behaviour as the verb dəən ‘walk’, in contrast from the noun phráʔ ‘monk’:
(4.9)| khon | thîi | dii |
| clf:person | rel | good |
| ‘the person who is good’ (verb-like property term, intransitive predicate) |
| Thai | TK | Post Reference Post2008: 344 |
Post (Reference Post2008) describes numerous further grammatical tests, including negation and aspect marking, showing that property concept words in Thai can be grouped together with words denoting actions rather than words denoting things.
Upon closer examination, Post shows that sub-classes of verbs can be distinguished from each other, and this can form the basis of an argument that adjectives are indeed a distinct class from (or at least a sub-class of) verbs in Thai. Post (Reference Post2008: 345) notes that a property concept like sǔuŋ ‘tall’ can enter into a comparative construction, while processes or actions like khít ‘think’ and dəən ‘walk’ cannot:
(4.11)| khon | níi | sǔuŋ | kwàa | khon | nân. |
| [clf:person | prx]NP1 | [tall]x | [more]CMPR | [clf:person | dst]NP2 |
| ‘This person is taller than that person.’ (verb-like property term) |
| Thai | TK | Post Reference Post2008: 345 |
(4.12)| *khon | níi | khít/dəən | kwàa | khon | nân. |
| [clf:person | prx]NP1 | [think/walk]x | [more]CMPR | [clf:person | dst]NP2 |
| *‘This person thinks/walks more than that person (does).’ (stative/active verbs) |
| Thai | TK | Post Reference Post2008: 345 |
Post gives further evidence showing how property concept words such as dii ‘good’ can be distinguished at a finer level of grain from action/event words such as dəən ‘walk’: this evidence includes distinct patterns of nominalization (Reference Post2008: 349), adverbialization (p. 353), and a specific pattern of reduplication with intensifying meaning (p. 364).
The case of Thai adjectives illustrates the typical situation in MSEA languages: property concept words will be like typical verbs in many respects, and in a few respects they will be distinct from those verbs. Depending on the details in the case of any specific language, and depending on a researcher’s inclination and perhaps theoretical persuasion, one may either conclude that there is no adjective class in the language (adjectives being basically verbs) or that adjectives form a distinct word class but have a lot in common with verbs.
Similar statements can be found for languages across MSEA. In Vietnam Mien, ‘adjectives share many properties of verbs’ (Doan and Mai Reference Doan and Mai1992: 89). For example, both adjectives and verbs take direct negation with the preposed negator mɤ̆j1:
And both adjectives and verbs can be directly modified by adverbs such as the preposed marker cɤ̆j5:
The issue of whether there really are adjectives in MSEA languages (and many other languages of the world) remains controversial, and can only be addressed seriously with detailed analysis of the full facts for each language. Such depth of analysis goes beyond our scope here, and indeed beyond the usual scope of the kinds of reference grammar works that we rely on in this book. In any case, for present purposes, we can safely say that adjectives are distinctly verb-like in this part of the world.
4.1.3 ‘Adpositions’ Are Often Nouns or Verbs
A third generalization about word classes in the MSEA area is a noted lack of a distinct word class that is widely observed elsewhere in the world, namely the distinct class of prepositions. In many MSEA languages, meanings of English prepositions such as to, with, by, from, on, under, through, and behind – members of a distinct form class in English and many other languages – are expressed either by verbs or nouns in specific grammatical constructions, as illustrated in the following examples:
(4.17)| Nws | cog | zaub | ntsuab | rau | hauv | vaj | lawm |
| 3sg | plant | vegetable | green | be.at | interior | garden | already |
| ‘She planted green leafy vegetables in her garden.’ |
| Hmong | HM | Clark Reference Clark and Bradley1989: 188 |
(4.20)| léih | deui | hàaih | hái | jēung | tói | hahbihn |
| 2sg | pair | shoes | be.at | clf | table | underside |
| ‘Your shoes are under the table.’ |
| Cantonese | SN | Matthews and Yip Reference Matthews and Yip1994: 117 |
While verbs and nouns carry a significant load in performing preposition-like functions, such as adding a peripheral argument (e.g., a goal of motion, or a location of an event or action) to a clause, this does not mean that MSEA languages have no true prepositions. For example, in Lao, the word duaj4 is neither a verb nor a noun, but functions as a dedicated preposition-like particle which can introduce a comitative or instrumental argument into a peripheral position in the clause. Here is an example:
(4.21)| man2 | kin3 | khaw5 | duaj4 | mùù2 |
| 3sg.b | eat | rice | with | hand |
| ‘He eats/ate rice with his hands.’ |
| Lao | TK |
This said, the meaning conveyed in (4.21) is more idiomatically conveyed by a multiple-verb construction, in which the verb saj4 ‘to use’ (or qaw3 ‘take’) takes as its complement a noun referring to the entity used as an instrument:
(4.22)| man2 | saj4 | mùù2 | kin3 | khaw5 |
| 3sg.b | use | hand | eat | rice |
| ‘He eats/ate rice with his hands.’ |
| Lao | TK |
Other MSEA languages similarly have few true prepositions, often having just one preposition with a semantically general locative meaning. In Kri, for example, the preposition taa takes locations as its complement; the word does not have other grammatical functions such as noun or verb:
(4.23)| hanq | vàt | phàsìì | taa | naaq |
| 3sg | spend | tax | loc | dem.ext |
| ‘He paid tax there.’ |
| Kri | AA | Enfield field notes (050729b:02.32) |
(4.24)| lôôh | taa | kuraaq | cawq | naaq |
| exit | loc | path | 1pl | dem.ext |
| ‘(It will) emerge at our path there.’ |
| Kri | AA | Enfield field notes (050729b:14.31) |
4.1.4 Verbs Can Act as ‘Adverbs’ without Derivation
MSEA languages tend not to have large classes of adverbs comparable to derived English words such as slowly, carefully, accidentally, and apparently. A frequent way to modify verbs in MSEA languages is simply to use another verb as its direct modifier, without any formal derivation. See Chapter 6 for examples, under the rubric of serial verb constructions.
4.2 Compounding
A compound consists of two open-class words joined together to form a single word. In English, for example, blackbird is a compound consisting of the two words black and bird. This is distinct from the phrase black bird in two important ways. The first concerns how it is pronounced. The compound blackbird takes stress only on the first syllable, while in the phrase black bird, the two component words take roughly equal stress. The second difference concerns meaning. The meaning of the phrase black bird is simply a combination of the meaning of the two component words – a black bird is a bird that is black – while a compound often has a meaning that, in the typical manner of morphological derivation, cannot be deduced from the sum of the parts: a blackbird is not any bird that is black (and indeed many blackbird species are not black).
Kratochvíl (Reference Kratochvíl1968: 73ff) provides a nuanced analysis of compounding in Mandarin Chinese. His analysis is useful as a model for MSEA languages in general. He identifies numerous categories of compound. For example, coordinate compounds resemble the structure of syntactic coordinate constructions. The meanings of the root forms and of the whole derived word are similar. For example:
(4.26)| jiāo ‘to establish contact’ + tōng ‘to go through’ → jiāotōng ‘communication’ |
| Mandarin | SN | Kratochvíl Reference Kratochvíl1968: 75 |
Subordinate compounds are of several types, including attribute-head, head-referent, head-modifier, head-measure, and reduplicated, as in the following examples, respectively (see Kratochvíl Reference Kratochvíl1968 76ff for details):
Jenny et al. (Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 38) give an overview of compounding in Austroasiatic languages. They note that, as with English examples like blackbird, morphological compounds are distinguished from syntactic phrases mainly by stress pattern and semantics (Jenny et al. Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 38; although they note on p. 41 that Khasi has overt morphological markers to distinguish compounds from phrases). Apart from stress patterns, another signal of compounding is sometimes observed in tone changes to a stem that is used in a compound. Here are some illustrative examples from White Hmong:
(4.32)| zaub | ntsim |
| vegetable | peppery |
| ‘peppery vegetable’ (syntactic phrase; meaning=sum of parts) |
| White Hmong | HM | Ratliff Reference Ratliff1992: 59 |
(4.33)| zaub-ntsig |
| vegetable-peppery |
| ‘peppery vegetable [as a particular plant or a pickled dish]’ (morphological compound; idiomatic meaning) |
| White Hmong | HM | Ratliff Reference Ratliff1992: 59 |
(4.34)| muaj | txiaj |
| have | money |
| ‘have money’ (syntactic phrase; meaning=sum of parts) |
| White Hmong | HM | Ratliff Reference Ratliff1992: 60 |
A frequent structure of compounds is the juxtaposition of noun and noun. The resultant meaning is often not predictable from the sum of the parts, but of course it is motivated by the meanings of the parts. Following are examples from across the MSEA language families:
Another common structure for compounds in MSEA languages is the combination of noun and verb/adjective, as illustrated in the following:
There are also many compounds consisting of verb and verb. As Jenny et al. (Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015) note, verbal compounds ‘are not always clearly distinguishable from multiple-verb predicates or serial verb constructions’ (Jenny et al. Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 40; cf. Chapter 6, below). These might combine verbs with two related meanings, as in the following example:
Or they might combine verbs that are synonyms, or near synonyms, as these examples show:
In this example from Phan Rang Cham, the compounding of two verbs yields a noun:
Certain quite specific compounds are shared across the area. For example, many MSEA languages lack a word meaning ‘parents’, but instead use a noun–noun compound of ‘mother’ and ‘father’. The relative order of the two parts of the compound varies. Compare the first and following examples here:
The examples shown above in this section only scratch the surface of the phenomena of compounding in MSEA languages. Like any process of morphological derivation, the details of compounding are textured and complex. There is a great array of underlying formal structures, and an even greater array of underlying semantic relations. The process of compounding shows variation from language to language in terms of its formal properties, its productivity, and its underlying semantic principles, if any can be detected. One of the defining characteristics of compounds is that their semantics are idiosyncratic. Most of the work of describing compounds, then, is a matter of careful lexicography, and so will tend to receive only superficial or tentative treatment in grammatical descriptions.
4.3 Reduplication
Reduplication is a linguistic process used in most if not all languages of the world. In an overview of reduplication in Austroasiatic languages, Jenny et al. (Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 42) identify two formal parameters for distinguishing between sub-types. One parameter concerns whether the reduplication is full or partial. The other concerns whether there is stem alternation. These formal parameters cross-cut various semantic functions. The distinction works for the other languages of the MSEA area as well.
In its simplest form, reduplication is simply repetition of a specified form, where this repetition conveys a conventional meaning. For example, in Thai, in one pattern of reduplication, when a noun is reduplicated, the output is a plural form of that noun:
(4.55)| dek~dek |
| child~rdp |
| ‘children’ |
| Thai | TK |
In Khmer, reduplication of a noun can signal a meaning of semantic generalization:
In Xia’ao Zhuang, reduplication of a numeral classifier means ‘each and every’:
Reduplication can signal a family of meanings relating to attenuation of the intensity of an action, or distribution of an ongoing action over time. In Taiwanese, for example, reduplicating a verb ‘suggests that the action is performed for a short unspecified amount of time’ (Lin Reference Lin2015: 214):
(4.58)| góa | beh | seng | khòaⁿ~khòaⁿ |
| 1sg | want | first | see~rdp |
| ‘I want to read this book a little (before I buy it).’ |
| Taiwanese | SN | Lin Reference Lin2015: 214 |
In Phan Rang Cham, full reduplication signals ‘some sort of distributive meaning’ (Thurgood Reference Thurgood, Adelaar and Himmelmann2005: 494), as shown in this example that includes vowel mutation in the reduplicated syllable:
And in Vietnam Mien, reduplication of an adjective serves to attenuate the intensity of the predicated property:
The semantic function of attenuation is also shown in this example from Muong:
The following examples of reduplication from Katu, Vietnamese, and Wa show the function of marking continuation or distribution of an action over time:
Reduplication in Eastern Kayah Li is not a word-formation mechanism, but has the meaning ‘also, too, either’. This form of reduplication is unusual in that it is ‘a simple matter of copying whatever syllable happens to be clause-final, regardless of either form-class or syntactic function’ (Solnit Reference Solnit1997: 52):
(4.65)| vɛ̄ | ma | ʔe | kʌ̄ | phɛ́ | thɛ́ | ja~ja |
| 1SG | be.so | eat | COM | simply | pig | flesh~RDP |
| ‘I only ate pork, too (as did he).’ |
| Eastern Kayah Li | TB | Solnit Reference Solnit1997: 52 |
To look firstly at cases of full reduplication and no stem alternation, we see some clear tendencies in meaning. Full reduplication in Austroasiatic languages often functions to modify predications, for example with attenuation of intensification, such as in the following examples:
Another widespread function of full reduplication with no stem alternation is ‘plurality or non-specificity’, both broadly construed. See the following examples:
Reduplication with stem alternation is widespread. Here are some examples of reduplication with stem alternation in a selection of Austroasiatic languages (the Vietnamese example showing stem alternations in tone, Khmer showing alternation in onset, and Bru and Mang showing alternations in vowel and/or rhyme):
In Jahai, the possibilities for stem alternation are numerous, with a single stem potentially undergoing several distinct alternations to mark several distinct meanings:
(4.76)| cɨp | ‘go’ | |
| cɨp~cɨp | continuative | full reduplication, no change |
| cip~cɨp | durative | full reduplication, with vowel alternation in first syllable |
| cp~cɨp | imperfective | full reduplication, with vowel deletion in first syllable |
| ca~cɨp | reciprocal | partial reduplication, with rime alternation in first syllable |
| Jahai | AA | Jenny et al. Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 44–45 | Burenhult Reference Burenhult2005: 94f |
In some cases that resemble reduplication, there are no attested corresponding base forms from which the apparent reduplications are derived. Just as we know in English what cranberry means without having a meaning for cran-, what looks like a reduplication may not have a known source stem in the language. Here are two examples:
Austroasiatic languages in particular show complex forms of elaborative reduplication. Kruspe (Reference Kruspe2004) describes a range of processes of derivational reduplication in Semelai. For example, in a form of partial reduplication called light syllable reduplication, everything in the stem up to the final coda is copied and affixed before the stem. Light syllable reduplication mostly applies to verbal roots with an intensification function. When applied to nominal roots, this conveys the meaning of plurality of the referent. The following examples show light syllable reduplication of verbs with the semantic effect of intensification and plurality, respectively:
(4.79)| ki=brkas | ʔilɔ~ʔilɔk |
| 3A=tie.together | intns~be.good |
| She tied (them) together really well. |
| Semelai | AA | Kruspe Reference Kruspe2004: 149 |
Another type of derivational reduplication in Semelai is coda copy. The coda consonant of the final syllable of a bisyllabic stem is attached to the coda position of the stem’s open penultimate syllable. When the stem is monosyllabic, its onset is also reduplicated to form a penultimate syllable, while the V slot is filled by an epenthetic rule. The functions of coda copy are diverse. The following example illustrates coda copy reduplication of a transitive verb, with the semantic effect of intransitiving the verb and conveying an imperfective aspect:
Coda copy reduplication in Semelai is also productive in word-class derivation. For example, a noun can be changed into a stative verb. In one pattern, nouns of body parts and clothing are converted to a verb meaning ‘to have N’:
(4.83)| sk-suk |
| rdp~hair/fur/feathers |
| ‘to have hair, fur, feathers’ |
| Semelai | AA | Kruspe Reference Kruspe2004: 151 |
These examples only scratch the surface of the complexities of Semelai morphological processes. See Kruspe (Reference Kruspe2004) for further details.
Let us now turn to the case of morphological processes in Vietnamese. This is worth doing for two reasons. First, thanks to Thompson’s excellent grammar of Vietnamese, we have an especially rich description of a single language system to draw on. Second, because Vietnamese has been cited since Sapir (Reference Sapir1921) as a canonical isolating language, we can take the opportunity here to temper the myth that there is effectively nothing in the way of productive morphology in this language. As we are about to see, this is not the case.
Thompson (Reference Thompson1987: 150ff) gives a detailed account of what he calls specializing derivatives in Vietnamese. Above, we have described these sorts of derivations as involving reduplication, or partial reduplication, with possible stem alternation. Kruspe’s approach, in her Reference Kruspe2004 description of Semelai, is to give formulae for constructing the appropriate output of these derivational patterns. Thompson conceives of the matter differently. He suggests that Vietnamese has ‘chameleon affixes’, which ‘have some kind of consistent phonetic resemblance to their bases’. Thompson’s idea is that a morpheme is added, but the phonological form of the morpheme is not fixed in advance, rather it surfaces in the same form as its immediate environment, in the same way that a chameleon changes colour to match its surroundings. Here we consider four types of chameleon affix:
1. Perfect chameleon affixes ‘have exactly the same phonetic shape as the bases with which they occur’ (p. 139); i.e., full reduplication without stem alternation.
2. Tonal chameleon affixes ‘differ from their bases only in tone’ (p. 139); i.e., full reduplication, with alternation of the stem only in the tone.
3. Riming chameleon affixes retain the rime of the stem (p. 140); i.e., reduplication in which the initial of the stem is changed.
4. Alliterative chameleon affixes retain the initial of the base (p. 140); i.e., reduplication in which the rime (everything but the initial) is changed.
Thompson (Reference Thompson1987: 150ff) goes in detail through a vast number of patterns and regularities of formation of these ‘affixes’ or reduplicative structures. Let us now consider just a sample of the great variety of types of reduplication that he identifies.
Among the perfect chameleon affixes, Thompson distinguishes a number of types including ‘distributives’ (based on nouns), ‘iteratives’ (based on verbs), and ‘attenuatives’ (based on adjectives), as in the following examples:
Another category of perfect chameleon affixes is ‘intensives’. These are formed from range of bases, as illustrated in the following examples:
A set of ‘emphatics’ are formed ‘with extremely diverse affixes’ (pp. 154ff). Following are six of the types that Thompson identifies (with the stem alternation in bold face, to indicate which part of the stem has been replaced with new material).
1. Emphatics with tonal affixes (p. 156):
2. Emphatics with vocalic affixes (p. 157); mostly ‘suffixing’ (i.e., infixing in the second syllable):
3. Emphatics with riming prefixes (p. 158); most common are b- and l-:
4. Emphatics with riming suffixes (pp. 158–159).
5. Emphatics with alliterative prefixes (pp. 159–60):
(4.95)| rắc~rối |
| rdp~be.mixed.up, tangled |
| ‘be complicated, intricate’ |
| Vietnamese | AA | Thompson Reference Thompson1987: 159 |
6. Emphatics with alliterative suffixes (p. 160ff). These are ‘by far the most numerous and diverse’:
Thompson provides over 300 examples, exemplifying dozens of patterns of stem-alternating reduplication (‘suffixes’ in his terminology). In terms of their productivity, these run the gamut from relatively productive (e.g., -a with over twenty examples given) to entirely ad hoc.
Thompson suggests that this system comes from a historical stage of the language in which the morphology was more transparently productive (p. 178):
[I]t may be argued that at an earlier stage of the language there was an elaborate morphological system, involving many types of chameleon affixes with more or less consistent meanings. These may or may not have been related historically to a variety of onomatopoetic and other single-morpheme forms of reduplicative type.
While we have concentrated so far in this section on Austroasiatic languages, we can note that these types of stem-alternating reduplications are also found in MSEA languages of other families that are often thought to lack productive morphological processes. An example is Lao. Table 4.2 shows examples of a form of expressive reduplication in Lao which takes a noun N and yields an expression meaning ‘Ns and things associated with them’. The reduplicated syllable is postposed, with stem alternation in which a back vowel is changed into a front vowel at the same height.
Table 4.2 Expressive reduplication in Lao (TK) with stem alternation, back vowel changed to front vowel
| Original form | Reduplicated form |
|---|
| cɔɔk5 ‘drinking glass’ | cɔɔk5 cɛɛk5 |
| khuq1’bucket’ | khuq1 khiq1 |
| pɯm4 ‘book’ | pɯm4 pəm4 |
| toq2 ‘table’ | toq2 teq2 |
| kɔɔng4 ‘camera’ | kɔɔng4 kɛɛng4 |
| pɔɔ3 ‘hessian’ | pɔɔ3 pɛɛ3 |
| kɯa3 ‘salt’ | kɯa3 kia3 |
| hua4 ‘fence’ | hua4 hia4 |
When the source stem does not feature a back vowel, a schwa may be used in the reduplicated syllable, as shown in Table 4.3.
Table 4.3 Expressive reduplication in Lao (TK), front vowel changed to schwa
| Original Form | Reduplicated Form |
|---|
| tip2 ‘rattan rice container’ | tip2 təp2 |
| piik5 ‘wing’ | piik5 pəək5 |
| pet2 ‘duck’ | pet2 pət2 |
| pheet4 ‘sex’ | pheet4 phəət4 |
| sek1 ‘cheque’ | sek1 sək1 |
| dɛɛk5 ‘eat (vulgar)’ | dɛɛk5 dək5 |
For further possible types of expressive reduplication in Lao, see Enfield (Reference Enfield2007).
4.4 Affixation
Many MSEA languages lack affixation in the usual sense of the term, as is expected given their isolating/analytic typological profile. Languages of the Austroasiatic language family are a notable exception. This exception is significant given that, as pointed out in Chapter 1, Austroasiatic languages make up nearly half of all languages found in the core MSEA area. In this section, we will draw mostly on the excellent overviews of Austroasiatic morphology presented by Jenny et al. (Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015) and Alves (Reference Alves, Enfield and Comrie2015), to give a taste of the forms and functions of affixation in the area.
Austroasiatic affixes are mostly prefixes and infixes. To different degrees across the Austroasiatic family, forms of affixation that appear to have been productive in ancestor languages lose their productivity in the course of historical development. At one end of the scale, morphology is clearly visible and productive in Aslian languages such as Semelai (Kruspe Reference Kruspe2004). At the other end of the scale, it is hardly visible at all in languages including Vietnamese, Chong, and Bunong (Jenny et al. Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 45). Somewhere in between, it is visible but less productive, and to an extent frozen, in languages including Mon and Khmer (Jenny et al. Reference Jenny, Weber, Weymuth, Jenny and Sidwell2015: 46).
Affixation in MSEA is overwhelmingly if not exclusively derivational. In this section, we survey a range of examples of affixation, illustrating some of its many functions and forms in MSEA.
4.4.1 Existential/Locative Marking
In languages of the Bahnaric and Katuic sub-branches of Austroasiatic, a nasal prefix marks existential/locative function on pronouns. Here are two examples:
4.4.2 Lexical Category Derivation
In many Austroasiatic languages, prefixes and infixes can be used to change the lexical category of a root, for example deriving nouns from verbs, or changing the valency of a verb (e.g., from intransitive to transitive). Here are cases of noun-deriving infixation from three branches of Austroasiatic, most of which derive clearly from a Proto-Austroasiatic -n- or -rn- infix (see e.g., Sidwell Reference Sidwell, Bowern, Evans and Miceli2008):
Similar derivations are made by prefixes in some Austroasiatic languages, as shown in the following examples:
The last three examples are fairly transparent in terms of their semantics. The following two examples from Car show derived meanings that are related to the meanings of their source stems, but that are not predictable from them:
4.4.3 Collective/Reciprocal Marking
In many Austroasiatic languages, a prefix marks a collective or reciprocal derivation on verbs, as shown in the following examples:
This example from Bru shows the same derivational meaning marked by an infix:
In many other languages of MSEA, collective or reciprocal marking involves a stand-alone marker. This marker may occur after the verb, as in Lao and Khmer:
(4.116)| khaw3 | hen3 | kan3 |
| 3pl.b | see | recip/coll |
| ‘They saw/see each other.’ |
| Lao | TK |
(4.117)| kee | cəŋ | khniə |
| 3pl.b | see | recip/coll |
| ‘They saw/see each other.’ |
| Khmer | AA |
Or it may occur before the verb, as in this Mandarin example:
(4.118)| tāmen hùxiāng kàndao le |
| they recp see asp |
| ‘They saw/see each other.’ |
| Mandarin | SN |
Another strategy is for the idea of reciprocity to be spelled out in a more explicit construction of the form ‘A V B, B V A’. This is idiomatic in Cantonese and Lao, as in the following examples:
(4.119)| Ngóh | béi-min | kéuih | kéuih | béi-min | ngóh |
| I | give-face | him | he | give-face | me |
| ‘He and I respect each other.’ |
| Cantonese | SN | Matthews and Yip Reference Matthews and Yip1994: 87 |
(4.120)| khòòj5 | tii3 | caw4 | caw4 | tii3 | khòòj5 |
| 1sg.p | hit | 2sg.p | 2sg.p | hit | 1sg.p |
| ‘You and I hit each other.’ |
| Lao | TK | Enfield Reference Enfield2007: 331 |
4.4.4 Causative
In many languages around the world, a causative form of a verb can be derived by a morphological process, which changes the meaning of a clause from ‘X Vs’ (e.g., ‘X dies’, or ‘X eats rice’) to, roughly, ‘Y causes X to V’ (e.g., ‘Y kills X’, or ‘Y feeds X rice’). In many Austroasiatic languages of MSEA, a causative form is derived using a verbal prefix. The precise form of this prefix varies across the languages, but there are strong formal commonalities: an initial bilabial plosive or bilabial nasal is followed by a vowel, usually a low front or central vowel. Here are some examples:
In some Austroasiatic languages, the causative derivation involves an infix: for example, an infixed m in Kui, an infixed vowel u in Old Mon, and an infixed a in Kri:
(4.127)| p<a>raang |
| <caus>cross |
| ‘take across’ |
| Kri | AA |
In another kind of derivation, the following two examples show prefixes deriving intransitive verbs from transitive verbs, where the erstwhile undergoer becomes the verb’s subject:
In many other MSEA languages, this valency alternation, in which the O argument of a transitive verb can be the S argument of an intransitive verb, is not explicitly marked at all. For example, in Lao:
(4.130)| man2 | tat2 | cia4 | lèèw4 |
| 3sg.b | cut | paper | prf |
| ‘She has cut the paper.’ |
| Lao | TK |
(4.131)| cia | tat2 | lèèw4 |
| paper | cut | prf |
| ‘The paper has been cut.’ |
| Lao | TK |
While valency-changing derivations are achieved by affixation in many Austroasiatic languages – and thus in the lion’s share of MSEA languages – and many other languages of the area, valency-increasing operations are done syntactically, in ways that are comparable with the dominant English strategy (e.g., I broke the cup → I caused the cup to break, He swept the house → I made him sweep the house). MSEA languages do not, however, mark subordinate verbs as non-finite, as in the English examples just given. Instead, MSEA languages realize this strategy via serial verb constructions (see Chapter 6, below).
4.4.5 Further Morphological Derivations
We close this section by noting some further functions of derivational morphology in Austroasiatic, in order to emphasize the breadth and array of morphology found in languages of the family that accounts for nearly half the languages in core MSEA.
Sedang features an ‘adversative’ prefix, as shown in the following example:
In Danau, a ‘gerundive’ ‘non-finite’ prefix forms an abstract noun from a verb:
Kơho has a nasal prefix that is added to demonstratives and other nominals to derive interrogatives. The prefix consists of a nasal stop whose place of articulation agrees with that of the stem-initial consonant. Here are two examples:
Similarly, in Mon, prefixation on demonstrative/interrogative stems derives pronouns and locative or manner adverbs. Here are some examples (illustrating accompanying morphophonological change in the prefix):
In Sedang, prefixation on numerals can derive measure words (using the prefix tə-) and ordinals (using the prefix mə-), as shown in these examples:
Finally, in Bugan, a number of lexical pairs suggest a range of lexical semantic alternations corresponding to alternations in vowel, tone, and initial consonant alternation, as illustrated in the following three examples, respectively:
4.5 Tone in Word Formation
It is not uncommon for speakers of tone languages to use pitch and other features of tonation in morphological and syntactic processes. This is widely observed in African languages. However, descriptions of MSEA languages have not widely reported any role of tone in morphology and syntax. This may mean that there is not much to observe. Or it may mean that with further work we will discover some patterns and functions that have not yet been noticed. Some existing work is suggestive of this second possibility. For example, we saw in the description of reduplication (‘chameleon affixation’) in Vietnamese, above, that tones are implicated in morphological processes: they can be targeted in reduplicative stem alternations.
The most important work in this area relating to MSEA languages is by Martha Ratliff, in her 1992 monograph Meaningful Tone. Ratliff examines the ways in which tone in White Hmong can play meaningful roles in the language, beyond simply serving to mark contrast between lexical items. Ratliff (Reference Ratliff1992) gives a detailed account of three of these roles that tone plays: (1) tone sandhi in compound formation, (2) defining lexical classes, and (3) as a component in the morphology of ideophones.
4.5.1 Tone Sandhi in Compound Formation
When compounds are formed, often there are phonetic/phonological markers of the relationship between the elements that are compounded, as noted for the role of stress in English compounds. In Hmong, the tone of an element in a compound can change due to sandhi rules. Ratliff notes that the sandhi form of a word, with its altered tone, can become disassociated from the trigger of the sandhi, and can become an independent word with a distinct tone from its original tone. The historical (Proto-Hmongic) and the sandhi forms may coexist with slight meaning differences, as shown in Table 4.4.
Table 4.4 Doublets in White Hmong (HM) arising from sandhi effects
| Original (when stand-alone) | Sandhi-affected (when in compound) |
|---|
| caj root [edible] | cag root [inedible] |
| hauv source; base; summit | hau head |
Some other MSEA languages show intriguing pairs of words with related meanings, differing only in tone. It is possible that these could have emerged from similar historical processes.
4.5.2 Regular Grammatical Distinctions
There are cases in Hmong where tone appears to act as if it were an affix that marks regular distinctions in grammatical meaning or form class. In West Hmongic (of which White Hmong is a member) and East Hmongic branches of Hmong-Mien, the distinction between dual and plural in the second-person pronouns is made solely by tone: tones that developed out of the Proto-Hmong-Mien A1 category signal ‘dual’, those from A2 signal ‘plural’ (see Ratliff Reference Ratliff1992: 22). Here are the examples in White Hmong:
White Hmong has a set of spatio-temporal words, almost all of which have the -m tone (low falling, checked). They are used as prepositions, and appear to be cognate with, and/or derived from, corresponding spatio-temporal nouns. Ratliff (Reference Ratliff1992) refers to them as denominal prepositions. Here are two examples:
In addition, a change from the -m tone of the above denominal prepositions to a -d tone derives demonstratives (Ratliff Reference Ratliff1992: 12). Here is an example:
(4.145)| nyob | sab | nraud |
| be | side | outside |
| ‘outside; on the other side’ |
| White Hmong | HM | Ratliff Reference Ratliff1992: 113 |
A number of male–female noun pairs are distinguished by tone alone, where the -g tone (low falling, breathy) occurs in masculine nouns, and the -m tone (low falling, checked) occurs in feminine nouns. Here are some examples:
(4.148)| yawm (txiv) |
| ‘maternal grandfather; older men in mother’s line’ |
| White Hmong | HM | Ratliff Reference Ratliff1992: 126 |
(4.149)| yawg |
| ‘paternal grandfather; older men in father’s line; man, sir’ |
| White Hmong | HM | Ratliff Reference Ratliff1992: 126 |
4.5.3 Forming Ideophones
In her detailed description and analysis of the system of ideophones in White Hmong, Ratliff (Reference Ratliff1992) describes a range of ways in which the ideophones of the language – many of which are reduplicative and disyllabic – feature alternations between the two syllables not just in consonants and vowels, but also in tones. For example, an ideophone duj-dig is derived from a stem dig ‘to be blind’. The ideophone is formed by prefixing a copy of the stem, changing the vowel to u, and changing the tone from -g (mid falling, breathy) to -j (high falling). Here is an example of the ideophone in its grammatical context:
(4.150)| maub | duj-dig |
| grope | ideo-blind |
| ‘to go along, feeling one’s way like a blind person’ |
| White Hmong | HM | Ratliff Reference Ratliff1992: 143 |
Another example involves the stem teev, a noun meaning ‘a drop (of liquid)’. Again, the ideophone is formed by prefixing a copy of the stem, changing the vowel to u, and changing the tone from -v (mid rising) to -j (high falling):
(4.151)| dej | nrog | tuj-teev |
| water | drip | ideo-drop |
| ‘the water drips drop by drop’ |
| White Hmong | HM | Ratliff Reference Ratliff1992: 144 |
These examples only scratch the surface of the workings of a highly complex and nuanced system. For a richly detailed description, both synchronic and diachronic, see Ratliff (Reference Ratliff1992).
