The Language of the Modhupur Mandi (Garo), Volume 1

Adverbial Affixes

(A)

The two obligatory constituents of verbs, verb bases and principal verb suffixes, were described in the previous chapter. This chapter will describe the remaining two constituents, the adverbial affixes that fit between the verb base and the principal verb suffix, and the terminal verb suffixes that come last of all. Each verb needs one, but only one, principal verb suffix, but a verb can have a whole string of adverbial affixes or it can have none. These affixes permit verbs to convey a great deal of specific meaning that is conveyed in English by separate words. They are a distinctive feature of the Garo language, and skillful speaking requires the ability to toss them off easily.

There are some limitations on which affixes can be used with which verb bases and in a few cases, particularly where adverbial affixes are contradictory in meaning, there are restrictions on which of them can be used with each other. Nevertheless, many adverbial affixes can be used in the same verb and they can be used in almost any combination. Unlike word order, however, where the language permits great freedom, the order of the adverbial affixes is almost completely fixed. By and large the more common adverbial affixes come late in the word, and these later affixes can generally be used most freely with the widest variety of verbs. A few adverbial affixes, in particular those that come near the end of the string, can be used with any verb in the language. Those that occur closer to the verb base are more idiosyncratic. Their meanings are not always as transparent as those that come later, and they are more restricted in the verb bases with whichPage  140they can be used. For the most part, adverbial affixes can be used freely with any of the principal verb suffixes and the meaning of the whole word generally follows in a transparent fashion from the meaning of its parts. Speakers freely construct new combinations of verb stems and adverbial affixes, and hearers must be able to reconstruct the overall meaning of words they have never heard before, just as people who listen to English must be able to understand the meanings of phrases they have never heard before. However, the derivational affixes that are found closestPage  141to the verb bases can form tight compounds with the verb base, and in this case the meaning of the resulting verb stems may not be fully predictable from the meaning of their parts.

Many adverbial affixes can be used in no other way than as adverbial affixes, but some can also be used as independent verb bases. The verb bases dil- 'lead, guide' and pil'- 'return', for example, can also be suffixed to other verbs as adverbial affixes. Indeed, so many forms can be used both as independent verb bases and as adverbial affixes, that derivation from verb bases seems the most likely origin of most adverbial affixes. However, some of the most frequently used adverbial affixes are never used as independent verb bases, and if they were originally derived from verb bases, any trace of that origin is now lost. Verb bases cannot now be productively converted into adverbial affixes, and it would be quite wrong to think of the building up of a sequence of adverbial affixes as simply a way of stringing verb bases together. To string ordinary verb bases together requires the first to have a subordinating suffix such as -e .

The language has scores of adverbial affixes. Those that are most common and most essential are described here, and a few examples are given of the less common types. Many others will be found in the word lists. They are described here in approximately descending order of their frequency and importance. This means that those that occur latest in the word (closest to the principal verb suffix), and those that are most consistent in their meaning, are generally listed early, while those that are less common and more idiosyncratic are described later. However a few important affixes come relatively early in the sequence and a learner will need these quite soon. The "A" list includes the most essential of the adverbial affixes. These can be used with nearly all verbs, and they have entirely clear meanings. Their freedom of use and high frequency entitles them to be called "inflectional" affixes. The affixes on the "B" list are a bit less essential but still very general in their meaning. The affixes on the the "C" list are less common. The examples given here are grouped by their meaning. Some describe shape and direction, some separation and togetherness, and some strength and completeness. The final list gives examples of more specialized "derivational" affixes. These are found closest to the verb base and they join with the base to form verb stems.

Adverbial Affixes: Inflectional

(A)

-ing-, -eng-, -ong- progressive. Mandis pronounce this affix as -ing- , the A'chik say -eng- , and speakers of some dialects of northern A'beng in the Garo Hills say -ong- . In whatever way this is pronounced, it is the sign of the progressive. Its resemblance to English -ing , of course, is pure coincidence. In fact the English and Mandi affixes are pronounced quite differently since the Mandi syllable ends with -ng , so i has its backed pronunciation. The progressive affix in Garo, as in English, indicates that the action or condition indicated by the verb is in progress. Bi-song kat-ing-a 'They are running'; Mu-a-ko dak-ing-jok? 'What have (you) been doing?' For English speakers, the meaning is rather transparent when used with words that translate English verbs, though it is used in a slightly wider range of circumstances than is the English progressive. For example, the Mandi sentence Ang-ko a-chak mang-cha-ing-a 'the dog is angry at me' has a progressive -ing although its natural translation into English does not. The Mandi progressive in this case indicates that the anger is in effect right now and that it is not necessarily permanent.

Since words with "adjective" meanings are intransitive verbs in Mandi, the progressive suffix can be used freely with words that have meanings such as 'big, small, hot' and so on, and the meanings of the progressive of these verbs are not so obvious to an English speaker. As with words with such meanings as 'run' or 'give', however, the progressive refers to the current and on-going condition. Ang-a cha'-ing-a 'I am eating (right now), I am currently eating'; Ang-a neng'-ing-a 'I am tired (right now), I am currently in the state of tiredness (but I am not necessarily in a permanent condition of tiredness)'; A-chak-pi'-sa chon-ing-a 'The baby dog is small right now, currently in the state of smallness' (but that is not its permanent condition); A-chak-pi-sa chon-a , without the progressive suffix, means 'The puppy is small', (quite likely permanently, the runt of the litter perhaps). -ing also sometimes indicates a developing condition. Ka'-sin-a means 'to be cold, of objects'. Ka'-sin-ing-a , with an added progressive, can be used to mean 'getting cold'.

If a verb has other adverbial affixes, -ing- is always the last, immediately before the principal verb suffix. -ing- is often used along with -ang 'in that direction', or -ba- 'in this direction' which are also adverbial affixes, and when used together as -ang-ing- or -ba-ing- they mean progressive development: dal'-ba-ing-jok 'getting big' (see below under -ang-, -ba- ). The progressive affix is very common and scores of examples are scattered throughout these chapters.Page  142

-ja- 'negative'. This affix also comes late in the sequence of adverbial affixes, but when both are present, -ja- always comes before the progressive -ing- . When used with most tense-aspect suffixes, the -ja- and the tense-aspect marker maintain their full forms: -ja-jok 'not any more', -ja-no-a 'will not', -ja-na-jok 'will not any more'. Where -ja- might be expected to occur with the present-neutral suffix, -a , however, it contracts. Instead of * -ja-a which would be expected as the non-contracted form, what is actually found is simply -ja . This can be regarded as the negative form of the present-neutral tense. It is one of a number of contractions that take place when vowels would otherwise face each other across a syllable boundary. If, however, a progressive -ing- follows the -ja- , the present-neutral -a is as securely present as any other tense-aspect suffix: cha'-ja-ing-a 'not eating'. -ja- negates the verb, and thereby the sentence in which it is used: kat-ja 'does not run', dal'-ja 'is not big', ken-ja-no-a 'will not be afraid', neng'-ja-jok 'not tired any longer', go-ja-na-jok 'will not throw any longer', kin-a-ja-ing-a 'is not hearing', chon'-ja-ing-jok 'no longer getting small'.

-ku- 'still, yet'. In positive sentences, the most natural translation of -ku- is 'still' but it is used more often in negative than in positive sentences, and the combination -ku-ja- means 'not yet'. -ku- always comes before -ja- . i'-ba-ku-a 'still comes', i'-ba-ku-ja 'has not yet come', i'-ba-ku-ing-a 'is still coming', -i'-ba-ku-ja-ing-a 'is not yet coming'. -ku-ja is a kind of semantic negative of -jok , although it is not the formal negative. -ja-jok is the formal negative of -jok but that means 'not any more'. This leaves a semantic gap, since -jok is a kind of 'perfect' tense that means 'has' or 'change of state'. The semantic negative of -jok might be expected to be 'has not', 'there has been no change of state'. This meaning is supplied by -ku-ja 'not yet' and this is often coupled with -jok in balanced questions, as though they are positive and negative opposites. The next example shows this coupling of -jok with -ku-ja . (Balanced questions will be described more fully in Chapter 14 "Restructuring").

The combinations *-ku-jok and *-ku-ja-jok are rare to nonexistent, even though both -ku-ja and -ja-jok are very common, and -ku-a is by no means rare. Indeed, it is difficult to know exactly what -ku-jok or -ku-ja-jok might mean if they did occur. Perhaps they would be something like 'still changed state' and 'still not doing it anymore'. I will let others worry about whether these should count as ungrammatical or simply meaningless.

-et-, -it-, -at- 'causative'. Both -et- and -it- occur in Mandi with no discernable difference in usage other than speaker preference. The A'chik form is -at- . In whatever way it is pronounced, the causative can be used freely with any sort of verb, whether transitive or intransitive. It precedes both the progressive and negative affixes: ron'-et-ing-a 'causing to give'. In some cases, -et- amounts to a transitivizer, turning an intransitive verb into a transitive one: mo-jim-a 'shake back and forth' (intransitive), mo-jim-et-a 'cause something to shake back and forth'; mik-brap-a 'be dazzled by a bright light', mik-brap-it-a 'dazzle (someone) with a bright light'.

Of course, the causative affix can be added to words that translate English adjectives as well as to words that translate English verbs: git-chak-et-a 'cause to be red, redden', dal'-et-a 'make large, enlarge'. Mandi has different intransitive and transitive verbs for 'boil': git-u-a intransitive and rit-a transitive; Chi git-u-ing-a 'The water is boiling'; Am-bi chik-ko rit-a-ming 'Grandmother boiled the water'. One woman, a fluent speaker, told me emphatically that it was perfectly possible to say git-u-it-a 'cause to boil' and that it meant the same thing as rit-a . She readily accepted both ang-a chi-ko git-u-it-a and ang-a chi-ko rit-a as meaning 'I boil the water'. Another fluent speaker told me just as decisively that git-u-et-a was utterly impossible, and that the only way to say that someone boils water is with rit-a . In fact, rit-a is so consistently used for the transitive sense of 'boil' that there is hardly a need for a consensus on whether or not git-u-it-a is possible. It is never necessary.

Except where a separate transitive word like rit-a may discourage or block a corresponding intransitive from taking a causative, the causative affix is highly productive. It can be used with almost any verb in the language, including transitive verbs: dok-et-a 'cause to hit, make (someone) hit', on'-et-a 'cause to give, make (someone) give'. I have occasionally heard double causatives—-two causative affixes used together: do-ga-ko chip-et-et-bo 'cause (someone) to close the door'. Mandis have accepted triple causatives that I have constructed and tested, but I have never caught one in actual use and I am a bit skeptical of their likelihood or genuine acceptability.

Sometimes, however, a causative is added to a transitive verb without much change in meaning. In this case the causative affix seems simply to emphasize the transitive nature of the verb: rik-a, rik-et-a 'drive away,Page  144chase away (especially animals)'; eng-a, eng-et-a 'untie'; grip-a, grip-et-a 'cover, put a cover on'. Go-a is a transitive verb meaning 'hit' and one speaker suggested to me that go-et-a with a causative means 'throw further or harder' than simple go-a . In the following example, also, the -et- in ke'-em-et-a simply strengthens the transitive nature of the verb, and makes the action more forceful:

Sometimes a causative verb can be used in two different ways. pin-ik-a is a transitive verb meaning 'point to, show'. pin-ik-et-a , with a causative affix, can be used with the same meaning, but it can also be used to mean 'make (someone) point'.

-ang- 'away' and-ba- 'in this direction' form a pair with opposite meanings. They are often used with verbs of motion, where -ang- indicates motion away from the position of the speaker while -ba- shows motion toward the speaker. mal-ang-a 'crawl away from here', mal-ba-a 'crawl in this direction', kat-ang-a 'run away', kat-ba-a 'run here', jro-ang-a 'swim away', bil-ba-a 'fly here'. Verbs such as mal-a 'crawl', bil-a 'fly', and kat-a 'run' can easily be used by themselves as long as there is no need to suggest anything about the direction of movement, though they can also be readily joined by one of the directional markers. I'-a is more consistently used with either -ang- or -ba- , yielding i'-ang-a 'go' and i'-ba-a 'come', though it is not impossible to use i'-a without a directional marker to mean simply 'move through space' without implying any particular direction.

The motion that is implied by one of these directional affixes can be somewhat metaphorical. Dal'-ba-ing-a 'getting large, growing' could describe a child who is growing up toward the height of the speaker. Dal'-ang-jok could describe a tree that has grown up and away from the height of the speaker. Even more abstractly, -ang- and -ba- can loose any sense of literal motion and simply convey the sense of 'becoming': Ai-ao, ang-a neng'-ba-ing-jok 'Oh dear, I'm getting tired, I've been getting tired'; jrang-ba-a 'grow light, as at dawn'. In this usage -ang and -ba become aspect markers that indicate change. Among other uses, they can indicate incipient action in a verb, a meaning that is more often lexicalized in English: mik-rak-ing-a 'is awake' has a progressive marker indicating an on-going state; mik-rak-ba-a 'wake up, come awake' shows that there is a change toward wakefulness. Mik-rak-ba-ing-a means '(he is in the process of) waking up'. -ang- and -ba- are very frequently used with the progressive -ing- and with -jok 'change of state':

These two directional markers always precede the other affixes that have already been described and they precede many that are still to be listed. Typically, they follow directly after the verb stem. Nevertheless, like the progressive, they are intimately involved in the semantics of tense and aspect. When used in the sense of 'becoming', -ang- and -ba- can be near synonyms, although even here they may be related to the state of the speaker. mik-rak-ba-ing-jok 'waking up' and chu-ang-ing-jok 'going to sleep' imply that the speaker is awake and that the sleeper's condition is approaching or moving away from that of the speaker. In their use as directional affixes -ang and -ba- are opposites, and in neither meaning can they occur together in the same verb.

-pil'-, -pil- 'return, reversed'. Garo has several other adverbial affixes in addition to -ang- and -ba- that indicate direction of movement. As an independent verb base pil'- means 'return, go back', and it carries that meaning into the verb to which it is affixed. As an independent verb it has a raka (glottal stop) but in accordance with the very general Garo rule, the raka is lost when it occurs as the second syllable of a word. It is retained when in third syllable position. i'-ba-pil'-a 'come back, return'; pang-pil-a 'reversed, upside down'; am'-pil-a 'turn away, turn around'; wal-pil-a , 'come back, return'; tang-pil-a 'come to life again after apparently being dead'.

In addition to -pil'- , A'chik speakers use the adverbial affix -tai- 'again'. I'-ba-tai-bo 'come again', Nam-gi-ja-gip-a kam-ko dak-tai-jok '(He) has done bad things again'. -tai- is rarely used in Mandi, and instead, -pil'- is extended in meaning to include the sense of 'again'. This means that, in Mandi, pil'- does not always imply a 'return' or an undoing of what went before, though it often does.

Most of the adverbial affixes that have been described up to now can easily be used together in the same verb: kat-ba-pil'-ku-ja-ing-a 'not yet running back here'.

-a-ri- 'just, merely, for no special reason'. -a-ri- is rarely used with a negative, but very frequently with the progressive. When followed immediately by the progressive -ing- the two affixes often contract to -a-ring- . Along with its central meaning of 'just, merely', it can sometimes suggest a certain surprise or rapidity. The same, or a homophonous, suffix with a similar meaning is also used as a subordinating suffix (see Chapter 6,Page  146"Verbs"). -a-ri is very common in Mandi. My impression is that it is less often used in A'chik:

Ak-bri si-a-ri-jok. Cls-four die-just-Prf 'Seven people simply died'.

A-gan-na man'-a-ring-a. speak-Inf can-just-Prog-Neut 'I can just talk'.

Gim-ik-ko-in cha'-a-ri-a. Ma-ming ji-a-ni dong-ja. all-Acc-Frg eat-just-Neut anything refusal be-Neg 'Just eats everything. There are no refusals' (expresses some surprise)'.

Ring-to da'-ring-to ring-a-ri-no-a.

drink-tasty not-drink-tasty drink-anyway-Fut 'Whether it tastes good to drink or not, (I) will drink (it)anyway.'

Bi-ming-ko ang-a-de gu-al-a-ri-a.

name-Acc I-Nomn-but forget-just-Neut 'I simply forgot (his) name'.

Un-i-ko-de i'-no-in dim-dak-in dong-tok-a-ri-ing-a. then here-Frg all-Frg be-at-all-just-Prog-Neut 'Then all of them are just staying here'.

Tek-ja, spok-e don-a-ri-a.

lock-Neg insert-Sub put-just-Neut '(It is) not locked, just left plugged in'.

Adverbial Affixes: General

(B)

The adverbial affixes of the next group are used a bit less often than those already listed, but they are still very common. They have somewhat more specific meanings than those in the first group, and they could be described as more "word-like" and less "grammatical". All of these except -be- 'very' precede -it, -ku, -ja-, and -ing- and all of them follow -ang- and -ba . The adverbial affixes in this set can generally be used productively with any verb base where the meaning is appropriate.

-be-, -bi- 'very, a lot'. Several adverbial affixes act as intensifiers. -be- is the most common and probably the least emphatic of them. jang-be-a 'very noisy', kat-be-a 'runs a great deal', at-ti mat-be-a 'the knife is very sharp', cha'-be-a 'eats a lot'. This is unusual among the adverbial affixesPage  147in being able to occur in more than a single fixed position. Most often it comes before -ja- as in dal'-be-ja 'not very big', but occasionally it follows -ja- instead, in which case it strongly emphasizes the negation: dal'-ja-be-a 'very much not big' which could be expected to be good deal smaller than merely dal'-be-ja 'not very big'. Dal'-ja-be-a does not come entirely easily, however, perhaps because of the danger of ambiguity.

-tok- 'all, everything'. Nik-tok-a 'see everything', kat-ang-tok-jok 'have all run away'. -tok- risks ambiguity since 'all' can refer either to the subject or the object of the verb: Ching-a bi-song-ko nik-tok-a could mean either 'We all see them' or 'We see them all'.

-grik- 'mutually, each other'. -grik- forms reciprocal verbs and it can be used productively with any semantically appropriate verb stem: a-gan-grik-a 'converse, talk to each other', sin-grik-a 'like each other', grong-grik-a 'meet each other'. -grik- is used with a few verb bases in ways that have been conventionalized so that the meaning is not fully predictable from the meanings of the parts. Dak-grik-a , with the verb base dak- 'do, make' might be expected to mean 'do to each other' but it has, instead, the specialized meaning of 'fight'.The following sentence describes two people meeting each other. It would be odd, perhaps ungrammatical, without the -grik- , or it might be understood to mean that the man and woman together met a third person:

Grong- 'meet' can be used transitively without the -grik- but with an accusative object instead: Me'-a-sa me'-chik-ko grong-a 'The man meets the woman'.

-srang- 'completely'. This is a stronger and more emphatic intensifier than -be- 'very': i'ang-srang-a 'completely gone, gone for good', a-gan-srang-ja-jok 'no longer says anything at all'. Like -be- , it has the unusual ability of being able to occur after as well as before -ja- and it is in fact a good deal more natural after -ja- than is -be- : Nam-ja-srang 'completely bad, terrible, awful', cha'-ja-srang 'eats absolutely nothing'. Since no principal verb suffix needs to follow -srang this has to be seen as a rather eccentric and idiomatic expression. It might even be proposed that -srang- counts as a separate word, different from the homophonous adverbial affix. In that case we would expect the principal verb suffix -a to have contracted with -ja- which would explain its apparent absence. However, nam-srang-a 'completely good' has the tense-aspect suffix -a after the -srang- , and -srang- is right where any proper adverbial affix belongs. * Nam-a-srang Page  148and * nam-srang are utterly impossible. The absence of -a from nam-ja-srang is an anomaly.

-cheng- 'first, before doing anything else': neng'-tak-cheng-bo 'rest first, (before doing anything else)'; ang-a cha'-cheng-no-a 'I will eat first'.

The same syllable, in related senses, is found in some other circumstances than as an adverbial affix. A'-ba-cheng-a 'begin' appears to include this affix, but a'-ba- does not exist as a separate verb base. Cheng-o is an adverb meaning 'in former times'. It usually has a locative -o , but the first syllable is clearly related to this adverbial affix. After the negative imperative, -cheng has the strange ability to occur after the tense-aspect marker -a , instead of before it, where proper adverbial affixes belongs: da'-dak-cheng-a or da'-dak-a-cheng 'don't do it first'; da'-chok-cheng-a or da'-chok-a-cheng 'don't write first'.

Other adverbial affixes cannot appear in this position, and perhaps this late appearing cheng has more to do with the locative cheng-o than with the adverbial affix.

-so- 'first, ahead, before someone else, wait'. While -cheng- means to do one thing before another thing, -so- means 'to do something before someone else does it': Cha'-so-a 'eat before someone else', cha'-cheng-a 'eat before doing something else', seng-so-a 'wait', i'-ang-so! 'go ahead!, go first!'

-kal- 'more than, comparative'. This is the most common comparative affix in Mandi, while -bat- (see just below) is more common in the A'chik dialects with which I am acquainted, but both affixes are widely understood: dal'-kal-a 'bigger', kat-kal-a 'runs more'.

-bat- 'more than, exceed, go across'. This is the usual comparative affix in the A'chik dialect that I am best acquainted with. It is equivalent in meaning and use to -kal- , but unlike -kal-, -bat- is related to, and presumably derived from, the independent verb base bat-a 'cross, go over, pass'. When used as an adverbial affix it can retain that meaning, though it can also be simply a comparative: ga'-bat-a 'step across, step over'; nam-bat-a 'better', cha'-bat-a 'eat more', in-di-sik-bat-a 'this much more'. -bat- is also clearly related to the postposition bat-e 'more than':

-dil- 'lead, causitive'. The main causative affix in Mandi is -et- or -it- , but a second affix, -dil- , sometimes acts as a causative as well. As an independent verb base, dil- means 'lead, guide, show how'. As an adverbial affix it often retains this meaning, but at times its meaning extends to causation: re'-ang-dil-a 'lead, guide'; pi'-sa-ko cha'-dil-a 'lead the child in eating, feed, cause to eat'; ken-dil-a 'frighten' ( ken-a 'be afraid'); ha-bu-dil-a 'give someone a bath' ( ha-bu-a 'take a bath'); dong-dil-a 'care for a baby or small child' (lit. 'lead in being'); dak-dil-a 'show how, lead' (lit. 'lead in doing'); nang-dil-a 'bump together', 'cause something to bump' ( nang-a 'bump into something', intransitive ). Perhaps -et- is becoming increasingly grammaticalized as a transitive marker, while -dil- is coming to take its place in a more literally causative sense.

-it- and -dil- may carry somewhat different meanings, with -it- being more directly causative. Si-it-a 'cause to die' implies specific killing. Si-dil-a is 'allow do die' as by neglect. A verb with both causatives is possible, though not common: nik-dil-et-a 'show where something can be seen'.

-man'- 'get, accomplish, manage to, finish, be able'. As an independent verb the central meaning of man'- is 'be able, can' and it carries a related meaning into its use as an adverbial affix. It can be used productively with most verb bases, but -man'- has merged with some verb bases into lexicalized verb stems with conventionalized meanings: ku'-man-a 'good at language, good at talking'; ku'-man-ja 'speak badly, stutter' ( ku'-sik 'mouth, language'); nik-man-a 'notice, discover' ( nik-a 'see'); bil-man-a 'have strength'; sok-man-ja 'fail to reach' ( sok-a 'arrive').

-bru- 'falsely, pretending'. in-bru-a 'tell wrongly'; jom-bru-a 'pretend to be sick'; dak-bru-a 'pretend, pretend to do'; i'-ba-bru-a 'come when not expected, at a bad time'; bi-ka so-bru-a 'pretend to be angry'; pin-bru-a 'cover too much, as hair covering the face, a blanket over one's head'; in-grik-bru-a 'quarrel in jest'; ra'-bru-a 'follow without belief, as religious rules'.

-e- 'while going'. This indicates movement, travel. While not terribly common it it can be used quite productively:

In the following sentence the -e- of don-e-no-a 'will go and put' is redundant, repeating the motion already indicated by re'-ang-e 'while taking'. The adverbial affix -e- is homophonous with the subordinating suffix, but its meaning and syntax are quite different:

Adverbial Affixes: Specialized

(C)

The adverbial affixes that have been described in the previous two sections are very common, and they can be used freely and productively with a wide variety of verbs. Mandi has scores of others that are more specialized, and that vary from semi-productive to no more than minimally productive. A few examples of the least productive kind of adverbial affixes are described in the final part of this section. These are the kind that deserve to be called "derivational affixes". They form somewhat tighter constructions with the verb bases to which they are attached than do those listed earlier. The sample given here is intended to suggest the range of meanings that they convey and the kinds of verb stems that they can form. Some indicatePage  151something about shape or direction of movement and others show something about spacing or strength, but their meanings are extremely varied. Many other examples are given in the lexical lists in the second volume.

Most of the bases to which these adverbial affixes are attached can act independently as verbs, and the affix simply refines the meaning of the base. A few of the bases are not verbs at all, however, but nouns or numeral classifiers. The nouns may originally have been incorporated objects, but they have become an integral part of the verb stems. ja'-rik-a 'follow (someone walking)', for example, is formed from ja'- 'foot, leg' and -rik- 'go after, follow'. ja'-rik-a is a completely conventional word and it would be impossible to use rik- by itself in this sense. Since it comes first, the ja'- might be considered to be the verb base although it is never used without an adverbial affix. In some cases, the verb base has no independent use at all, as a verb or anything else, but gains its meaning only in conjunction with an adverbial affix. The adverbial affixes of these verb stems do have identifiable meanings and they can be used in related sense with other verb bases. For example deng-gok-a means 'bow deeply'. -gok- has the meaning of 'bend the body' in a number of other verbs: ( geng-gok-a 'sleep with the knees drawn up', gok-gok-a 'bend over', bam-gok-a 'bow') but if deng- ever had a meaning of its own, separate from -gok- , that meaning is now lost. Most of these adverbial affixes are used in no other way except as affixes, but a few can also be used as independent verb bases with the same or a related meaning.

A considerable number of adverbial affixes are reduplicated in form, the same syllable occurring twice, sometimes with a slight change. In their reduplication, they resemble lexical adverbs, which are very often reduplicated. A good many adverbial affixes can occur both in reduplicated and nonreduplicated form. The reduplicated version sometimes implies more repetition than the nonreduplicated form, though there may be little difference in meaning. -tip- and -tip-tip both occur as adverbial affixes with the meaning of 'cover': rim'-tip-a 'cover with the hands'; wen-tip-tip-a 'wind around thoroughly'; ga'-jet-a, ga'-jet-jet-a 'hold tight with the feet or legs'.

Shapes and Directions . The adverbial affixes do not fall into obvious categories, and they contribute all sorts of meanings to their verbs. A few meanings are characteristic, however, and a considerable number of adverbial affixes indicate something about the shape of the object being talked about or the direction of its movement. The following examples are typical.

-chok- 'pointed'. ku'-chok-a 'pointed, of a pen, an animal's nose, the crescent moon'; ( ku'- 'mouth'); ro'-chok-a 'long and pointed' ( ro'-a 'long');Page  152 ran'-chok-a 'thin, scrawny, of people', (lit. 'dry and pointed').

-chrit- 'in slices, into long pieces, in stripes, striped'. sal-chrit-a 'striped, as cloth, or a tiger' ( sal-a 'pull, draw a line'), den'-chrit-a 'chop into long pieces' ( den'-a 'cut, chop'), mat-chrit-a 'cut slightly, scratch' ( mat-a 'cut, wound').

-ket- 'filled, squeezed tight'. gap-ket-a 'filled up tight, no more space, squeezed in' ( gap-a 'full'); dap-ket-a 'push wet mud into cracks', 'bury'; jot-ket-a 'insert, plug up' ( jot-a 'insert'); chang-ket-a 'choke on something'.

-on- 'down, downward'. chok-on-a 'get down, get off, as from a bus' ( chok-a 'jump'); ni-on-a 'look down' ( ni-a 'look'); sik-on-a 'push down' ( sik-a 'insert'); bil-on-a 'fly down' ( bil-a 'fly').

-pru- 'through, cut through'. rat-pru-a 'cut through' ( rat-a 'cut, slice'); ga'-pru-a 'step through' ( ga'-a 'step'); dok-pru-a 'knock a hole through' ( dok-a 'hit').

-rik- 'follow, along the way'. kin-a-rik-a 'listen to' ( kin-a- 'hear'); ja'-rik-a 'follow' ( ja'- 'leg, foot'); hai'-rik-a 'recognize' ( hai'-a 'know'); dong-rik-a 'stay behind' ( dong-a 'be at'); gi-sik ra'-rik-a 'remember' ( gi'-sik 'mind', ra'-a 'take').

-ro-, -ru- 'long'. mang-ro-a 'long in the body' ( mang- 'classifier for animals'); chang-ro-a 'tall, of people, trees'; ja'-ro-a 'long, of pant legs, shirt tails, for a long time' ( ja'- 'leg').

-wen-wen- 'in circles'. kat-wen-wen-a 'run in circles' ( kat-a 'run'); bil-wen-wen-a 'fly in circles' ( bil-a 'fly'). As an independent verb, wen-wen-a means 'wind around, wrap around'.

Separate, Together, Pieces . Other adverbial affixes indicate whether objects are close together or spread apart.

-chap- 'together'. ra'-chap-a 'take together, take something with something else' ( ra'-a 'take'); sik-chap-a 'put together' ( sik-a 'insert'); dong-chap-a 'add a second spouse' ( dong-a 'be at, dwell').

-drak- 'separate, into pieces'. bit-drak-a 'spread apart, as thatch, clothing'; chit-drak-a 'tear off' ( chit-a 'tear'); o-drak-a 'take off, of banana peel' ( o-a 'open'); den'-drak-a 'chop into pieces' ( den'-a 'chop').

-gat- 'add, cover, put on top'. so'-gat-a 'push into fire' ( so'-a 'burn' trans.); di'-gat-a 'lift onto, load' ( di'-a 'lift'); ra'-gat-a 'adopt a child' ( ra'-a 'take'). As an independent verb gat-a means 'load onto'

-ok- 'come apart, take apart'. sal-ok-a 'pull out' ( sal-a 'pull'); go'-ok-a 'come apart, come loose'; po'-ok-a 'take off, take out, of hats, corks'. As an independent verb base ok-a means 'pull out, extract'.

-go'-ok-, -gok- 'broken apart, separated'. dok-gok-a 'hit and separate, as to knock off a handle' ( dok-a 'hit'); rong'-gok-a 'break off a round thing,Page  153such as the bottom of a glass' ( rong-a 'classifier for small round things'); gil-gok-a 'peel, what happens to skin when irritated' ( bi-gil 'skin'). As an independent verb go'-ok-a means 'separate, come apart'.

-grang- 'spaced'. ap-grang-a 'widely spaced of basket pieces, made with open spaces' ( ap-dal'-a 'wide'); wa-grang-a 'dented, of a knife blade' ( wa- 'tooth'); ka-grang-a 'make with open spaces, of a fence' ( ka-a 'tie, build fence').

-pri- 'into pieces'. ak-pri-a 'tear into pieces' ( ak-a 'pick, pluck, as fruit'); go-pri-a 'scatter' ( go- 'throw'); den'-pri-a 'cut into pieces' ( den'-a 'cut, chop').

Strong, Fast, Complete, Constant . Many adverbial affixes indicate the manner in which an action is performed.

-chrak- 'forcefully, uncontrolled'. bi'-chrak-a 'break off, as branches in the wind' ( bi-a 'break'); ki'-chrak-a 'defecate vigorously' ( ki'-a 'defecate', used in A'chik as a verb but used only in compounds in Mandi); ken'-chrak-a 'comb vigorously, as to get out lice' ( ken'-a 'comb'); pik-chrak-a 'pull out strongly, as with roots, stones etc.' ( pik-a 'pull out').

-rak- 'strong, fast'. gong-rak-a 'fast'; mik-rak-a 'wake up', (lit. 'strong in the eyes'), ( mik-ron 'eye'); mang-rak-a 'healthy, strong of body' ( bi-mang 'body'); ding-rak-a 'strong of long slender things, like string' ( ding-a 'classifier for string and long things'). As an independent verb base rak-a means 'strong, firm, hard, difficult, expensive'.

-tat- 'forcefully'. dok-tat-a 'hit forcefully' ( dok-a 'hit'); sik-tat-a 'grab by force, as sons-in-law were once captured'; ni-tat-a 'stare' ( ni-a 'look').

-tip-tip- 'repeatedly, thoroughly'. ka-tip-tip-a 'tie thoroughly'; ba-ji-tip-tip-a 'thoroughly tangled'; chot-tip-tip-a 'tear to pieces, of string and long things'.

-srok- 'lightly, loosely, gently, partially, incompletely'. dok-srok-a 'hit lightly' ( dok-a 'hit'); ha-bu-srok-a 'bathe quickly or partially' ( ha-bu-a 'bathe'); a-gan-srok-a 'tell partially, incompletely' ( a-gan-a 'say, speak').

-min-ek- 'soften, squash, crush'. jot-min-ek-a 'poke and make soft, crush' ( jot-a 'insert, poke into'); ga'-min-ek-a 'step on and crush' ( ga'-a 'step'). As an independent verb, min-ek-a means 'tender, soft, ripe'.

Adverbial Affixes: Derivational

(C)

This final group of affixes includes those that are the most specialized of all. Each is used with a restricted set of verb bases, and the meanings of the resulting verb stems are less consistently predictable from the meaning of the parts than are the meanings of the affixes listed earlier.

-brok- 'tasteless, bland, rotten, faded'. so-brok-a 'rotten, not good to eat' ( so-a 'rot'); chi-brok-a 'tasteless, bland, insipid' ( chi 'water'); mik-brok-a 'fade, loose color, grow dull' ( mik- 'eye'). Mik- and chi are nouns rather than verb bases. They cannot be used as verbs without their derivational affixes, and the forms with the affixes are really compounds. A few similar examples will be found among the following verbs.

-chak- 'accept, receive'. an'-chak-a 'take into the body, be sick' ( an'- 'body'); wang-chak-a 'open mouth to receive food from another' ( wang-a 'open mouth wide'); ni-chak-a 'act as a midwife, one who sees and receives' ( ni-a 'look').

-chak- 'for another, support, help'. This is homophonous with -chak- 'accept, receive', but seems to have a sufficiently different meaning to count as a separate affix. dak-chak-a 'help' ( dak-a 'do'); hai'-chak-a 'be understanding, perceptive' ( hai'-a 'know'); a-gan-chak-a 'answer, reply' ( a-gan-a 'say'); ron'-chak-a 'give help' ( ron'-a 'give'); ka'-sa-chak-a 'sympathize' ( ka'-sa-a 'love').

-jip- 'with fanning, flapping'. grang-jip-a 'flap wings slowly' ( grang 'wing'); jak-jip-a 'beckon, wave' ( jak 'hand').

-nik- 'look upon, consider, regard'. chon-nik-a 'look down on, see as small' ( chon-a 'small'); dal'-nik-a 'look up to, respect, see as large' ( dal'-a 'big'); nom'-nik-a 'appear easy' ( nom'-a 'easy, soft'); nam-nik-a 'like' ( nam-a 'good'); seng'-nik-a 'show off' ( seng'-a 'bright, smart'). As an independent verb nik- means 'see'.

-srip- 'slurp, swirl in the mouth'. ku'-srip-a 'swirl water in mouth to rinse it' ( ku- 'mouth'); sal-srip-a 'suck, as on a straw, have a stuffy nose' ( sal-a 'pull'); ring-srip-a 'slurp, as tea' ( ring-a 'drink').

-to- 'good, tasty, beautiful, comfortable'. gan-to-a 'comfortable to wear' ( gan-a 'wear'); ni-to-a 'beautiful to look at' ( ni-a 'look at'); kin-a-to-a 'beautiful to hear' ( kin-a-a 'hear'); jak-kal-to-a 'good to use, comfortable to use' ( jak-kal-a 'use'). As an independent verb, to-a means 'taste good'.

-dik- 'painful, unpleasant, uncomfortable, ugly', the opposite of -to- . ni-dik-a 'ugly to see' ( ni-a 'look'); kin-a-dik-a 'ugly to hear' ( kin-a-a 'hear'); dong-dik-a 'uncomfortable, in a bad state' ( dong-a 'exist'); sa'-dik-a 'hurt, be painful' ( sa'-a 'sick, hurt').

Terminal Suffixes

(A)

A rather heterogeneous set of suffixes can follow the principal verb suffix. One, -ming , expresses past time or conditionality and it works closely with the tense-aspect markers so it counts as part of the tense system.Page  155Others ask questions, label speech as a quotation, and express doubt. A sentence does not require a terminal suffix, but two, or rarely three, are sometimes used together with the same verb, and their order is fixed. While not obligatory, terminal verb suffixes convey important meanings, and they are very frequent.

-ming 'Past, Conditional'. Except for its rare appearances with something other than a verb (see "Final Noun Suffixes and Terminal Suffixes" in Chapter 8, "Nouns"), -ming always follows directly after one of the four tense-aspect markers, -a, -jok, -no-a , or -na-jok . -ming- is very closely linked to the tense-aspect suffix, and depending upon which tense-aspect suffix it follows, it can be a simple past, a past perfect, or a conditional. Generally it indicates that the statement made is not true at the present time, although it may have been true sometime in the past, or it could be true in the future or under some stated conditions. -ming is not used in the A'chik dialect, where -chim and -a-ha divide its functions. For the Mandi dialect, the combinations of -ming with each of the four tense-aspect markers need to be considered separately.

-a-ming 'simple past'. Mandi does not have a simple past tense suffix that corresponds to A'chik -a-ha , and the morphologically more complex -a-ming takes its place. It is formed from present-neutral tense-aspect suffix, -a , together with -ming , and it simply indicates that something happened in the past. Whereas -jok describes a change of state in which the result of the change is still with us, -a-ming is most often used when the result of the action is no longer true: Bi-a kat-ang-jok 'He ran away (and is still gone)', Bi-a kat-ang-a-ming 'He ran away, he had run away (but may have come back)'.

Progressives and negatives can be used with -a-ming in a straightforward way:

-jok-ming 'past perfect'. This usually indicates that the event or situation described took place before some other event or situation:

The negative -ja-jok-ming has the sense of 'not any more', just as -ja-jok does when used without the -ming :

-no-a-ming 'conditional', 'would have'. This is formed with the future suffix, and it has a conditional meaning. It is often used in sentences where another clause has -o-de 'if'. Since an -o-de clause cannot include a tense marker, its temporal reference must be inferred from the -no-a-ming clause:

-na-jok-ming 'present or future conditional'. This is formed from the immediate future -na-jok and -ming .Like -no-a-ming it has a conditional sense, but it is used for present or future time. Like -no-a-ming , this is often used in a sentence that also has a clause with -o-de 'if', or it can imply the sense of 'if' even if it is not explicitly stated:

As with -ja-jok-ming , the negative of -na-jok-ming means 'not any longer', 'not any more':

-chim 'conditional', 'past' (A'chik). This suffix is rarely used in Mandi, although Mandi speakers often know it well enough to recognize it as characteristic of A'chik. It corresponds, in large part to -ming . It can be used with the A'chik tense and aspect markers and it can mean either past or conditional (or both) just as -ming can. However A'chik also has the principal verb suffix -a-ha which takes over some of the functions that -ming has in Mandi. In effect, A'chik -chim and -a-ha divide the work of Mandi -ming . -chim , more consistently than -ming , means that the action described was true in the past but is no longer true. -a-ha is a past tense that indicates nothing about the present situation. -ming also leaves open the present situation, but it is more likely than -a-ha to imply that the present condition has changed: re'-ang-a-ha (A'chik) 'he went' (present condition unspecified); re'-ang-a-chim (A'chik) 'he went' (but is no longer there); re'-ang-a ming (Mandi) 'he went' (and is likely not to be there any more but that is not certain); re'-ang-jok (both dialects) 'He has gone' (and has not returned).

Doubts, Questions, Quotes . The remaining terminal suffixes follow -ming if it is also present. They express a variety of meanings.

-kon 'perhaps, maybe, probably'. This very common suffix is used to hedge a statement. It can follow -a, -jok, no-a, -na-jok or -ming , but it cannot follow an imperative. Occasionally it can be heard suffixed to something other than a verb:

Bon'-ku-ja-kon '(It is) probably not finished yet'.

-ma 'question particle'. This provides the ordinary way to ask a yes-no question. It can follow any of the tense-aspect markers: i'-ang-no-a-ma? 'will (you) go?'; kat-jok-ma? 'has (he) run away?'. No syntactic adjustments need be made to a sentence when -ma is added, but it is often used with a question intonation in which the pitch rises at the end of the sentence, both on the -ma itself and on whatever immediately precedes it. Indeed a question intonation by itself is enough to ask a question that requires no more that a 'yes' or a 'no' in reply. The -ma makes it unambiguous:

-ma cannot be used in the same sentence as a question word. Compare the following two sentences. The first asks its question with -ma . The second uses ma'-ko 'what'. Both cannot be used together:

-ni 'question particle'. This is a near synonym of -ma but less common. Sometimes it suggests an expectation of agreement in the same way that tag questions do in English. When it is used with negative -ja- it carries a special rhetorical emphasis, having the sense of 'and was it not ever the case that. . .?' in order to suggest that 'it was indeed the case':

-mo 'question particle assuming agreement' (A'chik). For A'chik speakers, this is equivalent to the tag-questions of English. It anticipates agreement on the part of the listener. It's intonation, which starts quite high and drops decisively in pitch is remarkably similar to the intonation of English tag questions. -mo is never used by the Mandis of Modhupur:

-na 'someone said, I have heard that, quotative'. This suffix indicates that the sentence is a report of what someone has said, rather than a report of one's own observation. The -na is never obligatory, even when one is reporting another's speech, but it is a way of cautioning the hearer that the speaker cannot vouch for the truth of what he or she is reporting:

-ne 'courtesy form, please'. An imperative can be made more polite by adding a -ne to the -bo : A-song-bo-ne 'Please sit down'; Nap-bo-ne 'Please come in'. A request for permission is also made more polite with a -ne . Requests are often made with a verb in the infinitive followed by -ne . The infinitive -na makes it clear that it is a request, and the -ne makes it more polite: Ang-a at-chong-na-ne 'May I sit down?' -ne can also follow a verb with a future or immediate future tense-aspect suffix. Either the future or the immediate future is often used to announce one's intentions. AddingPage  161 -ne shows courtesy and implies that the listener can in some degree share in the decision: Re'-ang-na-jok-ne 'I will go now, okay?', Ang-a in'-o chok-no-a-ne 'I will write here, if I may'. The intonation of -ne usually has a slight rise followed by a sharp fall. It is quite the opposite of the yes-no question intonation, but instead like the intonation of the A'chik tag question -mo .

-da strengthens an imperative. It is forceful. It can be abrupt or rude, but it is not necessarily so: Cha'-bo-da 'Eat!' is urgent but not necessarily rude; Re'-ang-bo-da 'Go away!' is likely to be rude and even to be intended as rude. Intonation and tone of voice make the intended meaning unambiguous.

-ro and -ai

(C)

Two other suffixes show up at the end of the verb in the same way as terminal suffixes so they have to be counted among them, but they are considerably more specialized and less common than the others.

-ro . This suffix is much less essential than the terminal verb suffixes already described. It can be used to correct a previous statement or a presumption. The correction may be a mild one. If I had told someone incorrectly that I did not have a pen ( kol-om dong-ja ), I could correct myself by saying dong-a-ro '(I) have one after all.' -ro is sometimes used to give a positive answer to a negative question, but it can also correct presumptions that are only implicit in the situation. If one person makes a claim that second person denies, the denial can be marked with -ro . It can follow a statement, but not a question or an imperative. My impression is that some individuals use -ro much more than others. It may be that it is associated with particular regions or dialects, and that I have heard occasional speakers from areas where it is used, but I do not know what its distribution is. Most speakers I know use it quite sparingly, though everyone certainly knows and understands it:

-ai (C) . This suffix strengthens a statement, or like -ro , it offers a mild correction to a previous assumption. Like -ro , also, it is much less common than the previously described terminal suffixes. Unlike -ro it can be used with imperatives and with question-word questions, but it cannot be used with yes-no questions.

The most remarkable thing about -ai is the number of contractions that it forms with other suffixes. If said slowly and carefully, speakers can pronounce both the suffixes from which the contractions are formed, except that -a-ai and -ja-ai are difficult to elicit. In rapid speech the two original syllables reduce to one. This is very odd, because such contractions are not at all common in Mandi, or in other dialects of Garo, and -ai is by no means a high frequency suffix. One would expect contractions to affect the most frequent sequences, but sequences with -ai are certainly not highly frequent. The suffix -ai has neither an initial nor a final consonant, and it is possible that this phonological form encourages contractions, but neither the locative -o nor the subordinating -e are subject to contractions, although both are very much more common than -ai . -ai participates in the following contractions:

-a 'present-neutral', + -ai > -ai

-ja 'negative', + -ai > -jai

-na 'quotative' + -ai > -nai

-ming 'perfect' + -ai > mai

-bo 'imperative' + -ai > -bai

-ai follows many other affixes without contracting. With the exception of -ming , it does not contract after a syllable that ends in a consonant, and even after -ming contraction is less consistent than elsewhere. Thus the following do not contract: -jok-ai, -ing-ai, ing-jok-ai, -kon-ai, -na-jok-ai . Nor does -ai contract when following either form of the future tense-aspect suffix, -no-a or -no , even though these both end in a vowel.

-ai and its contractions almost always come at the very end of a sentence, though a postposed noun phrase can occasionally follow: Re'-ang-jok-ai, ang-a-de 'I, on the other hand really have gone'. The only way that -ai could be suffixed to anything other than a verb would be in an equational sentence that lacks a verb entirely: I-a bol-pang-ai 'This, indeed, is a tree'

-ai adds insistency to a statement or it offers a correction. Its use can be seen in the possible answers to the question kol-om dong-a-ma? 'Do (you) have a pen?' A neutral answer would simply be dong-ja '(I) don't have (one)'. If the question is insistent, a stronger answer would be dong-jai where the -ja- and -ai have been contracted to -jai , and wherePage  163the implication would be that the questioner really believes that the other person does have one, and the speaker wants to make a decisive denial. Of course if someone believed that you did not have a pen you could be equally insistent that you did have one by saying kol-om dong-ai 'I do too have a pen'.

Kat-ang-jok-ai. '(He) really did run off.'

Ma-ko dak-ing-ai? 'What in the world are (you) doing?'

Re'-ang-a-mai? 'Did (you) go?'

-ai can politely strengthen an imperative:

A-song-bo-ai, A-song-bai. 'Do sit down'.

It can be used with the negative and with combinations of other affixes:

Ok-kri-ku-jai. '(I am) not yet hungry!'

The quotative -na contracts with -ai as -nai . -ai strengthens an assertion, but -na attributes it to someone else, which reduces the speaker's responsibility for the accuracy of the statement. When used together as -nai , a firm assertion is made that the speaker really did hear another's claim, but the speaker still cannot attest to the truth of that claim. It remains hearsay:

Ok-ri-a-nai. '(He) really says (he) is hungry'

Bi-a a-chak-ko nik-jok-nai. 'It is definitely said that she saw the dog'

The terminal verb suffixes that have been discussed in this section are not quite so firmly and exclusively attached to verbs as are the preceding principal verb suffixes. Occasionally they turn up attached to a noun or to some other part of speech, but the vast majority of their appearances are with verbs. The use of terminal suffixes with nouns and other parts of speech will be described the section called "Final Noun Suffixes and Terminal Suffixes" in Chapter 9, "Nominals".