Is there any language where the past tense is the base form of a verb?

Question

The fictional language Flaidish has this feature. But I recently found out about a natural language (Mixtec) where the present isn't the base form of a verb, its the future tense.

I found this surprising. Flaidish is of course fictional, but I never thought that there would be a natlang where the simplest tense wasn't the present. Though in Mixtec's case, many verbs are identical in the present and future tense (though not all).

In most languages of course, the other tenses are at least as complex as the present tense, if not more so (note how in English, past tense is marked with -ed or a stem change, while the present is mostly just the bare infinite, plus MAYBE the third person singular ending -s).

Answer 1

First, it is important to be clear on what "most basic form" as described above covers. One notion is "structurally simplest", that is, "having the fewest added things". The other is "phonologically best for predicting other variants". Mixtec seems to qualify as an example of the future being "most basic" because (a) the future has no prefixes or suffixes, (b) certain lexically specified change initial /k/ to [ʃ], and other change tone, in the present tense.

In Logoori, there are 4 future tenses and 5 past tenses; there is no "simple present", there is a present progressive, and a habitual which is a construction build from a hard to gloss particle plus the progressive. The present progressive has a progressive aspectual suffix -a- (a-gʊ́r-áa "he is buying"), and the form does not actually mark the present (it marks the progressive, like English "talk-ing"), since that is how you form the future and past progressives as well.

All verb forms except the imperative have a subject prefix, many have a tense prefix, and all, including the imperative, have a final suffix, either /ɪ/, /i/, /aa/ or, in lieu one of the preceding, /a/. We can exclude the tenses with the special suffixes in the search for the most basic form, since /ɪ/, /i/, /aa/ are non-basic. The candidates are:

Indefinite future: a-ri-gʊ́r-á "he may buy"
Immediate future: a-ra-gʊr-a "he is about to buy"
Most immediate past: y-aa-kʊ-gʊr-a "he just bought"
Recent past: y-aa-ka-gʊr-a "he bought"
Remote past: y-áá-gʊ́r-a "he bought" (long ago)
Imperative: gʊr-a

The imperative wins the competition for having the least morphology. Apart from obvious prefixes, the indefinite future and remote past also have special tone melodies; the imperative does as well, though it is not manifested in the toneless root class that includes /gʊr/.

Since the imperative is usually classed as a mood rather than a tense, and assuming the question is about "which tense", we could strike the imperative from consideration. It then appears that we have one tense that adds two things (the immediate future: a subject prefix and -ra-), and four that add three things (subject prefix and two prefixes, or subject prefix plus one prefix and a tone pattern). The competition seems to favor the immediate future as "most basic" (simplest) – however, this tense, unlike all other forms, cannot be used in a negative clause. With that competitor disqualified, we have a tie between the indefinite future and any of the past tenses, in terms of how many added things they have.

I would favor using the tone-melody as a disqualifier, because tone inflection is really not just one thing, it's an abbreviation for a set of interrelated processes (deletion of the lexical tone, placing another tone somewhere). If we count tone melodies as "more than one", we now have just two forms, the most immediate past and the recent past.

Answer 2

Probably this confusion is familiar within Indo-European linguistics. If we use the concept 'root' instead of 'base' we will understand this issue more accurately. In Indo-European languages, as far as I know, there is this notion of a root as a lexical entity: Root{reason}, can form: reasoning, reasonable, etc. In other languages with non-concatenative morphology like Semitic (Hebrew, Arabic, Aramaic, etc) and Afro-Asiatic like Berber, things are different.

This latter class, especially Berber, has sometimes one simple AFFIX as a root (for example, a voiceless palato-alveolar sibilant alone means 'eat'). Sometimes Berber expresses future using the root itself without any morphological change. In relation to your question, observe the following data from Berber (spoken in Northern Morocco): Root{adr} 'demolish'

PAST: adr-x 'I demolished' the 'x' is 'I'
PRESENT: addr-x 'I'm demolishing'
FUTURE: ad- adr-x 'I will demolish'

If you can see the gemination of the intraconsonant in 'addr' deriving the present, but the PAST 'adr' is more similar to the root, i.e. more basic. The future also is marked, in this case, with 'ad' (will), becoming more complex. So Here the present is derived from the PAST not the other way around. Consequently, the PAST is more BASIC than other other forms.

Answer 3

You could say that in Proto-indoeuropean this might have been the case and there are indications of this in some later languages like Ancient Greek, however this is definitely debatable given the complex morphology and non-trivial rules of ablaut in PIE verbs, so take it more like "it is possible", rather than "PIE did this in a consistent manner".

We analyse the PIE word into its root, from which all other forms can be derived using extensive and complex set of rules.

The present tense is derived typically by expanding the root by infixed -e- (in sanskrit, this is called guṇa form). Sometimes it is expanded even further (e.g. n-infixation or reduplication) but let's not worry about it.

The aorist (basically something like perfective past, not dissimilar to English past simple) is derived by prefixing a so called augment e- (but in PIE, it is debatable whether the augment was really a prefix, a clitic or even a separate word) and the root remains in the non-expanded (zero-grade) form (of course, like with the present, there is a number of other affixes that can form aorist).

Examples from Ancient Greek:

λειπ-ω, λειπ-ει - I leave, he leaves

ε-λιπ-ο-ν, ε-λιπ-ε-ν - I left, he left

Answer 4

Tibetan is one if I understand Sir Charles Bell correctly. He says that when Tibs talk about verbs qua verbs and use them daily conversation the standard for verbs is the perfect tense. You can check this out by looking at Thumi dGongs Ter, a popular Tibetan verb reference. Bell's book is Colloquial Tibetan. As I understand it Tibetan has no infinitive.

Answer 5

In Cape Verdean Creole, a creole of Portuguese, the root form of an action verb is used for the perfective aspect. I've read that the related Guinea-Bissau Creole is similar. And for action verbs, the perfective means it has already happened, similar to past tense. This root form usually sounds similar to the present tense, non-perfective form of a related verb in Portuguese (after applying regular sound changes). However it's not as simple as saying "root form = past tense", because:

1. Stative verbs are handled differently depending on the dialect, generally not using the root form in past tense situations
2. The root form is also used as the infinitive and and imperative forms
3. There are other past tense verb forms which are not the root verb

Despite these caveats, it's very very common to use the root form for verbs where the past tense would be used in English, or the preterite in Portuguese.

Wikipedia says that the Bajan Creole of Barbados, a creole of English, also uses the root form for the past tense, and similarly it looks like the root forms generally sound similar to present tense verbs in English.

Answer 6

It’s not exactly the ‘base’ form as such, but it is true in Semitic languages that the ‘most basic form’ of a verb is past-like or perfective.

As mentioned in Tsutsu’s answer, Semitic verbs employ non-concatenative morphology. In practice, this means in a nutshell that the (abstract) root of a verb consists of a purely consonantal skeleton; triconsonantal roots are the most common. Surface forms are then derived by applying various template-like patterns to this skeleton, a process which involves (1) geminating or reduplicating one or more of the consonants, and (2) injecting vowels before, between and after the consonants. In addition to these ‘templates’, surface forms often also include actual prefixes and suffixes.

These patterns can be fairly complex, but it is generally the case that the least complex form found in the paradigm is the third person singular masculine in the past (perfective/non-present) tense.

For an illustrative example consider this paradigm of the Classical Arabic root k-t-b: the 3sm past form is kataba. Underlyingly, the template is 1a2a3 (using numbers for the three root consonants), and the final -a is the personal ending for the third person singular masculine; but in practice, the surface form is equivalent to simply adding an /a/ after each consonant in the root, which is about as simple as you can get. Compare this with the third person dual masculine present indicative form yaktubāni, which uses the template 12u3, a prefix ya- and the personal ending -āni.

As a result of this, the 3sm past form is use in (almost?) all Semitic languages as the ‘base’ form that you look up in dictionaries and use to talk about verbs-as-verbs, similar to how the infinitive is used in most European languages. You cannot predictably derive the whole paradigm from this form (you need the consonantal skeleton plus often some extra information to do that), but it is the closest thing to a ‘base’ form that exists.

Answer 7

Surgut Khanty is a clear example of this; in Surgut Khanty the past tense is unmarked whereas the present tense is marked with the affix -ᴧ- (which represents the lateral fricative which in IPA is [ɬ]. For example, see this paper (page 34).