The maps in this and the following chapter show the distribution of suppletion in verb paradigms according to tense-aspect, imperative and hortative moods, and verbal number. Suppletion is defined as the phenomenon whereby regular semantic relations are encoded by unpredictable formal patterns. Cases where the paradigmatically related forms share some phonological material are examples of weak suppletion, as in English buy vs. bought, while cases with no shared phonological material are instances of strong suppletion, as in English go vs. went (Dressler 1985). These types reflect two ends of a continuum rather than an either-or opposition. Mel’čuk (1994) also introduces the criterion of uniqueness, i.e. in order to be classified as suppletive, the alternation shown between two paradigmatically related forms should be unique, in that no other two forms in the language show exactly the same morphophonological alternation, as in Spanish digo ‘say.pres.1.sg’ vs. dije ‘say.pret.1.sg’. The maps concentrate on strong and unique cases.
Formal irregularity is determined by synchronic criteria. In contrast to Rudes (1980), diachronic considerations such as the historical origin of the suppletive forms and their etymologies are not taken into account here when defining the similarity or dissimilarity of paradigmatically related forms.
In the literature (Anderson 1985, Carstairs-McCarthy 1992, Carstairs-McCarthy 1994, Mel’čuk 1994) it is customary to make a distinction between suppletion of stems, as in English go ‘go.pres’ vs. went ‘go.pst’, and suppletion of affixes, as in English boy-s ‘boy-pl’ vs. ox-en ‘ox-pl’, where the distribution of the plural affixes is not phonologically conditioned. The distribution of stem suppletion is presented on the maps below.
It is also customary to distinguish between inflectional and derivational suppletion. The maps exemplify both kinds. Tense and aspect as well as imperative are semantic categories associated with verbs in an inflectional way. Verbal number (or pluractional) is a category that rarely becomes inflectional (Mithun 1988b, Corbett 2000); it is associated with verbs in a derivational and even lexical manner. Some scholars, such as Mithun, Mel’čuk (Mel’čuk 1994) and Corbett, do not accept the suppletive status of verbal pairs expressing verbal number and rather see them as separate lexical items. However, the languages where such verb pairs are found exhibit a great deal of variation as regards the obligatoriness of their use as well as their function, so that a clear-cut decision about their status in the grammar and lexicon of the languages where they occur is often difficult. No distinction between inflectional and derivational suppletion is made for these maps.
Suppletion is generally considered to be rather unsystematic and accidental, hence difficult to explain or downright uninteresting. However, several studies (Bybee 1985, Markey 1985, Deshpande 1992, Aski 1995, Fertig 1998, Veselinova 2003) have shown that this assumption is not correct. The historical development of suppletive forms provides ample evidence that they are semantically motivated (Deshpande 1992) and also well integrated in the inflectional systems of their languages (Aski 1995). The synchronic data mapped below present evidence that it is not only possible to discern patterns of suppletion according to grammatical categories, but that these patterns can be further correlated with particular groups of lexemes and show rather different areal and genealogical distributions.
Map 79A shows the distribution of suppletion according to tense and aspect categories. It is accompanied by a supplementary map (see §3) where the distribution of suppletive forms in the imperative and hortative moods is presented.
| Go to map | ||
|---|---|---|
| Value | Representation | |
 |
Suppletion according to tense | 36 |
 |
Suppletion according to aspect | 10 |
 |
Suppletion in both tense and aspect | 24 |
 |
No suppletion in tense or aspect | 123 |
| Total: | 193 | |
English is an example of a tense-suppletion language, as in go vs. went. Suppletive forms may be used to mark one tense distinction only, as in English, or to mark several tense distinctions, as in Alamblak (Sepik; Papua New Guinea).
(1) Alamblak (Bruce 1984: 146): ‘go’
kit-wë-r
go-pres-m.sg
(yi)riah-r
go.fut-m.sg
yifi-r
go.hod-m.sg
r-i-ë-r
hest-go-hest-m.sg
yi-më-r
go-rem-m.sg
The second value, which covers languages where suppletive forms are used to express mainly aspectual distinctions, is illustrated by Oneida (Iroquoian; New York State and southern Ontario).
(2) Oneida (Karin Michelson, p.c.): ‘say’
kātuheʔ
1.sg.say.impf
wakǝ
1.sg.say.stative/perf
waʔkīluʔ
1.sg.say.pfv
Suppletive forms may be used to make the general imperfective vs. perfective distinction, as in (2) above, or to express finer-grained aspectual distinctions, as in Chalcatongo Mixtec (Oto-Manguean; Oaxaca, Mexico).
(3) Chalcatongo Mixtec (Macaulay 1996: 170): ‘come’
nbíí
come.hab
bèì
come.prog
kii
come.unrealized
na-kii
pfv-come
The third group of languages, which shows suppletion in both tense and aspect, includes languages where verbs coding both of these categories are encountered, as in Lezgian (Daghestanian; eastern Caucasus). It also covers languages with more complex verbs where the stem alternation is determined by both temporal and aspectual distinctions in one and the same paradigm, as in Russian below.
(4) Lezgian (Haspelmath 1993: 135-136): ‘be’ and ‘do’
a.
ja
be.pres
b.
awú-na
do-aor
tir
be.pst
ijí-zwa
do-impf
(5) Russian (Maria Koptjevskaja-Tamm p.c.): ‘come’
ja prixož-u
1sg come.pres-1sg
ja prixod-il
1sg come.impf-pst.m
ja priše-l
1sg come.pfv-pst.m
ja prid-u
1sg come.fut.pfv-1sg
ja budu prixod-it’
1sg be.fut-1sg come.impf-inf
The number of verbs per language showing tense-aspect suppletion tends to be very restricted, ranging from one single verb to ten at the most (more than three suppletive verbs are observed typically in the third group). The verb meanings encountered with all three kinds of tense-aspect suppletion are (in decreasing order of frequency) commonly used verbs such as the different kinds of copula (if it exists), ‘go/come’, ‘say/speak’, ‘see’, ‘give/take’, ‘do’, ‘eat’ and ‘die’. Other verb meanings, occurring in the third group only and often in very few languages, are ‘get’, ‘hear’, ‘stay’, ‘throw’, ‘put’, ‘call’, ‘become cold’, ‘become’, ‘bear’, ‘drink’.
Rudes (1980: 1) puts forth a claim that has rarely been questioned, namely, that suppletive verbs follow the same pattern in one the same language, tense/aspect/mode, on the one hand and person and number, on the other. However, the data collected here challenge this view, as is shown by the map of suppletive imperatives and hortatives. The latter phenomenon is illustrated by Egyptian Arabic.
(6) Egyptian Arabic (Mitchell 1962: 97): ‘come’
ana geet
I come.pfv.1.sg
ana aagi
I come.impf.1.sg
taʕaali
come.imp.f.sg
taʕaalu
come.imp.pl
| Go to map | ||
|---|---|---|
| Value | Representation | |
 |
A regular and a suppletive form alternate | 8 |
|
Imperative | 29 |
|
Hortative | 2 |
|
Imperative and Hortative | 1 |
|
None (= no suppletive imperatives reported in the reference material) | 153 |
| Total: | 193 | |
Unlike tense-aspect suppletion, suppletive forms in the imperative and hortative occur mainly with two verb meanings: ‘go/come’ and ‘give’. Other verb meanings in this group are verbs used for negative commands as well as ‘sit down’ and ‘be’, but the latter are not very frequent. Suppletive imperatives and suppletive forms marking tense-aspect distinctions appear as different phenomena both synchronically (see areal distribution in §4) and diachronically. Suppletive imperatives tend to result from incorporating verbs restricted to the imperative function into the richer paradigms of semantically similar verbs, or from borrowing in situations of intensive language contact. Neither of these pathways is particularly common for the development of suppletive forms according to tense and aspect.
Tense-aspect suppletion shows clear areal patterns. Its mainstay is Europe and the western parts of Asia, with the South Asian subcontinent as its eastern limit. The phenomenon is competely absent in Southeast Asia. Other hotbeds are found in Papua New Guinea and Mesoamerica. Tense-aspect suppletion is much more weakly represented in Africa, North and South America and Australia.
The distribution of the different kinds of tense-aspect suppletion shows areal patterns as well. In Europe and western Asia, the dominant types are tense and tense-aspect suppletion.
The use of suppletive forms to mark tense distinctions is the dominant type of tense-aspect suppletion in Papua New Guinea.
Languages with suppletive verbs following aspectual distinctions only are markedly absent in western Eurasia. This feature appears mainly concentrated in Mesoamerica with languages from the Oto-Manguean family. Some Southern Uto-Aztecan languages in this area show verbs with this feature as well as verbs displaying both tense and aspect suppletion. The few instances of tense-aspect suppletion in other parts of the North American continent are of the aspectual kind as well, occurring with Slave (Athabaskan; Northwest Territories, Canada) and Oneida.
In South America, the marking of tense distinctions by suppletive forms is found with all three Carib languages on the map as well as with the only Barbacoan language represented here; the use of suppletive forms to mark aspectual distinctions only is observed in Ika (Chibchan; Colombia).
In Australia, tense-aspect suppletion is observed with three languages: Maung, Mara and Nunggubuyu, all in the northern part of the continent, and all non-Pama-Nyungan.
Suppletion according to tense is represented by a limited number of languages in sub-Saharan Africa. It occurs in three Nilo-Saharan languages, Turkana (Nilotic; Kenya and Uganda) Kanuri (Western Saharan; Lake Chad region) and Fur (Sudan), as well as in two non-Bantu languages of the Niger-Congo family, Ewe (Kwa; Togo and Ghana) and Supyire (Gur; Mali); the latter is the only sample language in Africa where aspectual suppletive pairs occur as well. The other instances of tense suppletion in Africa occur in Harar Oromo (East Cushitic; Ethiopia) and Khoekhoe (Central Khoisan; Namibia).
Unlike tense-aspect suppletion, suppletive imperatives are hardly used in Europe except for Irish, Russian, and the languages of the Caucasus and the Balkans. Family-wise, they are present in all Uralic languages mapped here, which apart from Hungarian do not show tense-aspect suppletion. The stronghold of suppletive imperatives appears to be Africa: they are observed both with Afro-Asiatic languages north of the Sahara and with languages from all major language families in sub-Saharan Africa, the eastern parts in particular. They are also found in Mesoamerica in Jakaltek (Mayan; Guatemala) and Rama (Chibchan; Nicaragua), both languages where tense-aspect suppletion is not observed.