Indo European copula

General features

This verb has two basic meanings. In a less marked context it is a simple copula (I'm tired; That's a shame!), a function which in non-Indo-European languages can be expressed quite differently. In a more heavily marked context it expresses existence (I think therefore I am); the dividing line between these is not always easy to draw. Some languages have shared these functions between several verbs: Irish, Spanish and Persian all have multiple equivalents of to be, making a variety of distinctions. Many Indo-European languages also use the conjugations of the verb "to be" as an auxiliary for the formation of compound (periphrastic) tenses (I'm working; I was bitten). Other functions vary from language to language. For example, although in its basic meanings, to be is a stative verb, English puts it to work as a dynamic verb in fixed collocations (You are being very annoying).

The copula is the most irregular verb in many Indo-European languages. This is partly because it is more frequently used than any other, and partly because Proto-Indo-European offered more than one verb suitable for use in these functions, with the result that the daughter languages, in different ways, have tended to form suppletive verb paradigms. This article describes the way in which the irregular forms have developed from a series of roots.

*h1es-

The root *h₁es- was certainly already a copula in Proto-Indo-European. The e-grade (see Indo-European ablaut) is found in such forms as English is, Irish is, German ist, Latin est, Sanskrit asti, while the zero grade produces forms beginning with /s/, like German sind, Latin sumus, Vedic Sanskrit smas, etc. In PIE, *h₁es- was an athematic verb in -mi; that is, the first person singular was *h₁esmi; this inflection survives in English am, Persian am, Sanskrit asmi, Old Church Slavonic есмь (esmĭ), etc.

This verb is generally reconstructed for Proto-Indo-European thus:

*bʰuH-

The root *bʰuH- or *bʰuh₂- (which did not have ablaut variations in the protolanguage) probably meant 'to grow', but also 'to become'. This is the source of the English infinitive be and participle been (Germanic participles have the suffix in -an). Also, for example, the Scottish Gaelic "future" tense bithidh; the Irish imperative bí, past bhí and future beidh; and the Slavic infinitive, etc. for example Russian быть (byt’). PIE *b^h became Latin /f/, hence the Latin future participle futūrus and perfect fuī; Latin fīō 'I become' is also from this root, as is the Greek verb φύω (phúō), from which physics and physical are derived.

This verb can be reconstructed as follows:

*h2wes-

The root *h₂wes- may originally have meant "to live". The e-grade is present in the German participle gewesen, the o-grade (*wos-) survives in English and Old High German was, while the lengthened e-grade (*wēs-) gives us English were. (The Germanic forms with /r/ result from grammatischer Wechsel.) See Germanic strong verb: Class 5.

*h1er-

This has been claimed as the origin of the Old Norse and later Scandinavian languages' present stem: Old Norse em, ert, er, erum, eruð, eru; the second person forms of which were borrowed into English as art and are. It has also been seen as the origin of the Latin imperfect (eram, eras, erat) and future tenses (ero, eris, erit).

However, other authorities link these forms with *h₁es- and assume grammatischer Wechsel (/s/→/r/), although this is not normally found in the present stem. Donald Ringe argues that the copula was sometimes unaccented in Pre-Proto-Germanic, which would have then triggered the voicing under Verner's law. He explains the Germanic first person singular form *immi as such, deriving it from earlier *ezmi, since -zm-, but not -sm-, was assimilated to -mm- in Germanic (for which other evidence exists as well). Furthermore, the third person plural form *sindi (from PIE *h₁sénti) shows that this word, too, was unaccented. If the accent had been preserved, it would have become *sinþi, but that form is not found in any Germanic language. In this view, it is likely that stressed and unstressed varieties of the copula (with corresponding voiceless and voiced fricatives) existed side by side in Germanic, and the involvement of a separate root *h₁er- is unnecessary.

*steh2-

The root *(s)teh₂- meant "to stand". From this root comes the present stem of the so-called "substantive verb" in Irish and Scottish Gaelic, tá and tha respectively. On the absence of the initial s- in Celtic, see Indo-European s-mobile.

In Latin, stō, stare retained the meaning "to stand", until local forms of Vulgar Latin began to use it as a copula in certain circumstances. Today, this survives in that several Romance languages (Galician-Portuguese, Spanish, Catalan) use it as one of their two copulae, and there is also a Romance tendency for a past participle derived from *steh₂- to replace the original one of the copula (this occurs in French, Italian and the main dialects of Catalan). See also Romance copula.

Although in Dutch, this verb retains its primary meaning of "stand", it is also used in an auxiliary-like function that only has a secondary meaning of "standing". For example, in ik sta te koken ("I stand (while) cooking"), the emphasis is on the act of cooking, and the implication of standing while doing so is only secondary, and may even be completely irrelevant. German "stehen" also exhibits a similar secondary meaning as a copula in certain locative phrases. For example, "Das Auto steht da" ("the car is over there" lit. "the car stands there"). When it is not a full copula (it can normally only be used as an auxiliary with another verb), it does have shades of meaning that resemble that of the Italian sto cucinando ("I am cooking"). The intransitive verbs zitten ("to sit"), liggen ("to lie") and lopen ("to walk/run") are used in similar ways.

In Swedish, which usually lacks gerund forms, the corresponding stå is often used similarly, along with sitta ("to sit"), ligga ("to lie") and gå ("to walk").

Hittite

The Hittite verb "to be" is derived from the Indo-European root *h₁es-.

Vedic Sanskrit

The Vedic Sanskrit verb as (to be) is derived from the Indo-European root *h₁es-.

bhū (to be) is derived from Indo-European *b^huH-.

Persian

With regard to the function of the verb ‘to be’ as a copula, the most conspicuous feature of Modern Persian language is the evolution of an existential be, hast (exists), out of ast (is). In fact, when studying the forms and functions of ‘to be’, one might find certain characteristics specific to Persian that is worth pondering upon— i.e. even without considering the diachronic evolution of Modern Persian language and its relation to Ancient Iranian languages (such as Old Persian and Avestan) whose usage of the verb ‘to be’ seems more close to Sanskrit.
Paradoxically, despite the fact that Persian is apparently the only Indo-European language that has created an existential be out of the copula, it has simultaneously made an extreme use of the latter to produce a general paradigm for conjugating all Persian verbs.
Historically speaking, like most of Indo-European languages that make use of suppletive roots to denote ‘to be’, Persian integrates Proto-Indo-Eroupean (PIE) verbs *h₁es- (to be) and *b^huH (to grow> to become> to be). Hence, while Persian (Pers) infinitive būdan (to be) < PIE *b^huH forms the past stem of the verb (e.g. Pers būd- ‘was’) or acts as an auxiliary verb in formation of pluperfect of other verbs, its present tense is solely based on the derivatives of PIE *h₁es-. It is, in fact, from the declension of PIE *h₁es- (to be) that six present stems have been created and assigned to the 1st, 2nd, and 3rd person singular and plural to act as the present-tense conjugation of Pers būdan (to be), as shown in the following table.
(NB 1. Modern Persian nouns and pronouns have no grammatical gender; thus, all pronouns are neutral, e.g. the language does not distinguish between ‘he’ or ‘she’.
NB 2. In the conventional transliteration of Iranian languages: [x] is to be pronounced as ch in Bach or Loch Ness; [ā] as a in far; [ī] as ee in deed; [ū] as oo in food, [š] as sh in she, č as ch in chair, [ž] as j in déjà vu)

As an example, in the following sentences, the present forms of the verb 'to be' are used as copulas or predicates:
______________ "man doxtar-e to am; īn barādar-e man ast (a or e); to pedar-e man ī "
______________ (‘I am your daughter; this is my brother; thou art my father.’)
Furthermore, as endings added to the stem of the verbs, these declensional forms have been grammatized to shape a general paradigm for the grammatical conjugation of all other verbs; as if any of these endings was once an auxiliary verb which evolved into an enclitic. Ironically, this generalized conjugational paradigm is also applied to the past tense of the very verb būdan (e.g. būd-am = literally: *‘(I) am was’→ ‘I was’; similarly būd-īm→ ‘we were’ etc.)
(NB. An old declensional suffix -ad is still used instead of ast for 3rd person singular in present-tense and subjunctive conjugation of Persian verbs).
However, what is linguistically notable, is the emergence of an existential be out of the copula, viz hast (exists) out of ast '(is). The evolution of this exceptional form, might go back to ancient Iranian languages, where ast could have two variants (cf. Avestan which has both as- and has- <PIE *h₁es- ‘be’). In the next phase, what we may call a pseudo-verb appeared, vis. the verb هستن hastan (to exist) has been analogically evolved from hast (exists) and has been conjugated like any other Persian verb (e.g. hast-am = literally: *‘(I) am existence’→ ‘I exist’).
Nevertheless, created out of a wrong re-analysis of hast (exists), the restrictions of this pseudo-verb would show up as soon as one tries to either find its past-tense and/or negate it. Then, hastan (to exist) will simply turn into the familiar past-tense or negation of the aforementioned verb būdan (e.g. būd-am ‘I was’ vs. na-būd-am ‘I was not’) which acts just as a copula and its existential meaning has to be inferred like English ‘to be’. We may therefore deduce that neither hast (exists) nor its derived pseudo-verb hastan (to exist) can be negated at all, unless reduced to copulas! In other words, we do not have *ni-hast (it does not exist), but *ni-ast> nīst (it is not) (despite the fact that there was once such a verb as (a)nihistan in Middle Persian (Pahlavi)). It is also worth mentioning that in the early stages of the Classic (Modern) Persian, ni-am was the only way to say ‘I am not’. However, following a secondary analogy (similar to hast [exists] → hastan [to exist]) a negative pseudo-verb nīstan (not to be) was invented. Therefore, the mirror image of the pseudo-verb hastan is not a negation of the existential verb, but just a simple negative copula formed from ast (is) → *ni-ast→ nīst (is not)> nīstan ‘not to be’ which, like hastan, has only present-tense (e.g. nīst-am ‘I am not’).
Finally, going back to the verb būdan, there is no doubt that it can only act as a copula, even if we look at its other forms which, as a heritage of the ancient Persian, has provided such moods as optative bād [<bovād ‘May it be!’], (wish to be), the potential (mood of probability) bov-ad (might be), and perhaps subjunctive/ conditional/ imperative bāš.
(NB. However, bāš may be from another PIE root *wes- cf. English was).
In this regards, we may notice that the third person forms of bāš whose infinitive bāšīdan is conjugated to produce conditional, or substantial mood, has a tendency to replace ast (is), although considered wrong by educated/cultured Persians (e.g. ū mādar mī-bāš-ad [She is (a) mother.] instead of simply ū mādar ast).
To summarize: As mentioned in the General Feature, the most significant point is that Persian grammar differentiates the marked or substantive form هست hast from copula است ast. Hast is used to express existence while ast expresses predication. Of course in colloquial Persian, hastan can be also used for emphasis to express predication in the present tense. In the past tense, the verb بودن būdan covers both meanings.

Strictly speaking, hastan is only a theoretical infinitive, not lexical. Therefore, būdan functions as the actual infinitive., used, for example, in past-tense as būd (was) and in present-tense modal bovad ‘it may be’. The circumstantial form of būdan is باش bāš, used for the imperative and subjunctive, e.g. اگر باشد agar bāsh-ad 'if it is'.

The forms ast (is) and hast (exists) have been derived from the Indo-European root *h₁es-, where būd (was) is derived from Indo-European *b^huH-.

Greek

The Ancient Greek verb eimi (I am) is derived from the Indo-European root *h₁es-.

Note also that the participles are based on the full-grade stem ἐσ- in Homeric, according to Smyth.

Italic languages

Except for Latin, the older Italic languages are very scarcely attested, but we have in Oscan set (they are), fiiet (they become), fufans (they have been) and fust (he will be), and in Umbrian sent (they are). This section will explain Latin, and the Romance languages that have evolved from it.

In Spanish, Catalan, Galician-Portuguese and to a lesser extent, Italian there are two parallel paradigms, ser/èsser/essere from Latin esse "to be" on one hand, and estar/stare from Latin stare, "to stand" on the other.

For simplicity, the table below has only the full conjugation of the present tense, and the first-person singular forms of some other tenses.

In several modern Romance languages, the perfect is a compound tense formed with the participle as in English, but the old Latin perfect survives as a commonly used preterite in Spanish and Portuguese, and as a literary "past historic" in French, Italian and Catalan.

There is a tendency for a past participle derived from stare (or more specifically its supine, statum) to replace that of the main copula derived from esse. For example, the French participle été comes from statum.

For further information, see the main Romance copula article.

Germanic languages

Proto-Germanic retained the dual, but only in the first and second person.

Old English kept the verbs wesan and bēon separate throughout the present stem, though it is not clear that they made the kind of consistent distinction in usage that we find, for example in Spanish. In the preterite, however, the paradigms fell together. Old English has no participle for this verb.

The plural forms in Modern Swedish (indicated in brackets) were in common use in formal written language until the mid-20th century, but are now no longer in use except in deliberately archaising texts. The preterite subjunctive is also increasingly being replaced by the indicative.

Slavic languages

Baltic languages

In Lithuanian, the paradigm būnu, būni, būna, etc. is not considered archaic or dialectal but rather a special use of the verb būti, to be, mostly used to describe repeated actions or states, or habits.

Celtic languages

In the Celtic languages there is a distinction between the so-called substantive verb, used when the predicate is an adjective phrase or prepositional phrase, and the so-called copula, used when the predicate is a noun.

The conjugation of the Old Irish and Middle Welsh verbs is as follows:

The forms of the Old Irish present tense of the substantive verb, as well as Welsh taw, come from the PIE root *stā-. The other forms are from the roots *es- and *bhū-. Welsh mae originally meant "here is" (cf. yma 'here').

Irish and Scottish Gaelic

In modern Gaelic, person inflections have almost disappeared, but the negative and interrogative are marked by distinctive forms. In Irish, particularly in the south, person inflections are still very common for the tá/bhí series.

The verb bí

The copula

Gaelic (bh)eil and Irish (bh)fuil are from Old Irish fuil, originally an imperative meaning "see!" (PIE root *wel-, also in Welsh gweled, Germanic wlitu- "appearance", and Latin voltus "face"), then coming to mean "here is" (cf. French voici < vois ci and voilà < vois là), later becoming a suppletive dependent form of at-tá. Gaelic robh and Modern Irish raibh are from the perfective particle ro (ry in Welsh) plus ba (lenited after ro).

Modern Welsh

The present tense in particular shows a split between the North and the South. Though the situation is undoubtedly more complicated, King (2003) notes the following variations in the present tense as spoken (not as written according to the standard orthography):

Note that, for example, the spoken first person singular dw i'n is a contraction of the formal written yr ydwyf fi yn . The Welsh F /v/ is the fricative analogue of the nasal /m/, the PIE suffix consonsant for the first person singular.

Bod also has a conditional, for which there are two stems. The bas- stem is more common in the North, and the bydd- stem is more common in the South:

Armenian

The Classical Armenian present tense derives from PIE *h₁es- (cf. sg. h₁esmi, h₁essi, h₁esti; 3rd pl. h₁s-énti).

Albanian

The Albanian copula shows two distinct roots. The present jam ‘I am’ is an athematic root stem built from PIE *h₁es-. The imperfect continues the PIE imperfect of the same root but was rebuilt based on the 3rd person singular and plural. The perfect, on the other hand, comes from the thematic aorist of PIE *kʷelh₁- ‘turn’ (cf. Ancient Greek épleto ‘he turned’, Armenian eɫew ‘he became’, Old Irish cloïd ‘turns back, defeats’). Analogical or otherwise indirect reflexes are italicized below.