To be published in:
Kim, Ronald et al. (eds.): Diachrony and Suppletion. Conference held in Prague, September 2015.
Hamburg, Baar-Verlag (Studies in Historical Comparative Linguistics).
MORE TALES OF TWO COPULAS – THE COPULA
SYSTEMS OF WESTERN EUROPEAN LANGUAGES
FROM A TYPOLOGICAL AND DIACHRONIC
PERSPECTIVE
Britta Irslinger, Freiburg
Copulas are cross-linguistically widespread. The grammatical inventory of
languages may comprise one or more than one type. In the latter case, the
copulas differ in their functional scope (Pustet 2003, 45). Two-copula systems
can be found in several neighbouring Western European languages (table 1) and
have been examined as convergence phenomena in the context of linguistic
contact and linguistic substrata.
Romance Spanish ser estar
Catalan ser estar
Portuguese ser estar
Old French estre ester
Basque Western dialects izan egon
Insular Celtic Irish is tá
Insular Celtic Middle Welsh yw/ys byð
Germanic Old English is bið
Table 1: Western European languages with two-copula-systems
The Middle Welsh1 and Old English systems listed at the bottom of the table
have been analysed as Celtic influence on Germanic or Brittonic influence on
English first by Keller (1925, 56-60) and later on by modern scholars such as
Schumacher (2007/2009), Lutz (2010, 231-35) and Trudgill 2011.
Wagner (1959, 177f.) and Pokorny (1959, 156, 1962, 133), on the other hand,
compared the copula systems of Spanish, Portuguese and Irish with Berber
(Wagner also added Bantu) and concluded that the two-copula systems of Irish
and Romance emerged from an Afro-Asiatic substratum.
The substratum hypothesis of Vennemann (2010, 381) is more complex. He
assumes that “Vasconic”, a predecessor of Modern Basque once spoken in
Western and North-western Europe, had a two-copula system that influenced
1
The copulas is/iu vs. bid are already present in fragmentarily attested Old Welsh (c. 800-1050),
which is contemporary to Old English (Schrijver 2011, 68-70).
2 BRITTA IRSLINGER
Continental Celtic. From there, the feature passed on to Insular Celtic and Old
English on the British Isles and to North-western Romance on the Continent.
Contact with one-copula languages lead to the loss of the feature in later stages
of these languages, such as Middle English, Middle French and the Eastern
Basque dialects. In contrast to previous scholars, he discusses both sets of copula
systems together, claiming that the Middle Welsh and Old English copula
system and the Romance, Basque and Irish one both ultimately emerged from
the same linguistic substratum.
Vennemann’s substratum hypotheses, which he constantly developed over the
last decades, have received ample and appropriate criticism, both from Indo-
Europeanists and Vascologists.2
Nevertheless, Trudgill (2011, 13), whose primary interest lies in the alleged
Brittonic influence on Old English, considers Vennemann’s scenario as a
hypothetical albeit plausible background:
“... if there was a two-copula system in North-western Romance, it is not clear how
confidently we could ascribe the presence of this system to Celtic (and thus perhaps to
Vasconic) influence, but the case would be strong.”
He further argues that, although two or multiple copulas are cross-linguistically
frequent, the grammatico-semantic distinction, i.e. habitual vs. non-habitual,
found in languages of different origins is unusual and “does seem to signal some
kind of contact-based explanation” (p. 5). The process in question would be the
“additive complexification” of the Western IE languages by long-term contact
and bilingualism (p. 6).
However, these far-reaching conclusions are based on a number of assumptions
and premises that have not yet been discussed in detail. Research on the
diachronic development of the languages in question and on the functions of the
respective copulas has remained largely unnoticed.
This article will address the following topics: section 1 will discuss copulas in
typological research as well as suppletion within the copula paradigms of IE
languages. Sections 2 to 8 will examine the copula systems of the languages in
question from a synchronic and diachronic perspective. The findings will then
be discussed in section 9 and summarized in section 10.
2
See among others Trask (1995, 68ff.), Villar / Prósper (2005, 397ff., 413), Baldi / Cuzzolin 2006
and the articles in Udolph (ed.) 2013.
SUPPLETION AND THE COPULA 3
1. Crosslinguistic and diachronic background
1.1 Copulas in typological research
Before comparing the individual languages, it will be useful to briefly consider
the semantic relations of the copula as established by typological research.
Pustet (2003, 5) gives the following definition:
“A copula is a linguistic element which co-occurs with certain lexemes in certain languages
when they function as predicate nucleus. A copula does not add any semantic content to the
predicate phrase it is contained in.”
Table (2) is based, with modifications, on Dixon (2010, 159) and Pustet (2003,
28-33) and lists a number of semantic relations coded by copula constructions
in most of the languages discussed here.
relation type of copula copula copula copula complement
complement subject (predicate)
1 identity • definite NP This is John.
He is the teacher.
• substantivized adjective It is the red one.
• complement clause The idea was that John should lead.
2 classification/ • indefinite NP He is a teacher.
attribution • predicative adjective This man is clever.
oblique case predicates
2.1 possession • possessive phrase This book is John’s.
2.2 comitative • NP marked by He is with her.
adposition or affix
2.3 benefactive/ --"-- This is for John’s birthday.
recipient present
2.4 origin, source --"-- He is from Australia.
2.5 position / • NP marked by The tree is in the garden.
locative adposition or affix
• locational adverb over there.
3 temporal • temporal predicate It is eight o'clock.
4 existential • special construction There is coffee in the kitchen.
predicates • special verb Es gibt Kaffee in der Küche.
Table 2: Semantic relations in copula constructions (Dixon 2010, 159, Pustet 2003, 28-33)
The copula links a copula subject and a copula complement or predicate to
express identity (1). The copula complement is either a definite noun phrase, a
substantivized adjective or a complement clause. Another semantic relation is
classification, which is also called attribution or ascription, where the copula
complement is an indefinite noun phrase or an adjective (2). The copula
complement may be an oblique case predicate expressed by an NP or a
4 BRITTA IRSLINGER
prepositional phrase. The list of possible meanings given here is not exhaustive.
While Pustet (2003, 33) lists temporal predicates (3) as their own category, one
might consider them as a subtype of identificational predicates. Further, copulas
frequently occur in existential predicates, which may however not conform to
the usual structure of subject, copula and complement (4). Examples include
constructions like Engl. there is/are or German es gibt with another (transitive)
verb. According to Pustet it is not clear whether these constructions are really
devoid of meaning and thus qualify as copulas.
In locative oblique case predicates (2.5), English uses the copula with a locative
NP or a locational adverb, while other languages employ positional verbs. In
German, even though Der Baum ist im Garten is grammatical, the use of the
verb stehen ‘to stand’ would be preferable (Der Baum steht im Garten) – in
certain contexts at least. Depending on the entity in question, other position
verbs are used such as liegen ‘to lie’ or the more unspecific sich befinden ‘to be
situated, to be located’ (ex. 1).
(1) The tree is in the garden. Der Baum ist im Garten. Der Baum steht im Garten.
The book is on the table. Das Buch ist auf dem Tisch. Das Buch liegt auf dem Tisch.
The chair is in the yard. Der Stuhl ist im Hof. Der Stuhl befindet sich im Hof.
It is evident that these verbs have the potential to become grammaticalised as a
copula and thus enter the paradigm of the verb BE. However, it is debatable
whether they are always semantically empty as required by the definition cited
above. If they still add meaning to the phrase in which they are contained, they
are to be classed as semi-copulas or quasi-copulas (Pustet 2003, 5).
1.2 Copulas in PIE
Crucial to the argument is Vennemann’s (2010, 389) premise that “Proto-
Germanic, as indeed Proto-Indo-European, had only a single copula”. However,
there is not much evidence for this claim. According to Stassen (1997, 97), Indo-
European is the best example of “multi-rooted copulas”. The root *h1es- ‘to be,
exist’, which is present in all IE languages, is restricted to the present tense
paradigms in most languages.3 Only in Hittite it is not suppletive. The paradigms
of most Indo-European languages are composed of two or more verbal roots, cf.
the overview in Kölligan (2007, 125-127).
*h1es- + *kʷelh1- ‘to turn’: this is found in Armenian present 3sg ê : aorist ełew
and in Albanian present 3sg është : aorist qe (< *e-kʷl̥ h1-e/o-). In Greek the aorist
ἔπλετο ‘became, was, took place’ can be found as an archaism and was replaced
3
See Lühr (2015, 71-86) for the inflection of the present indicative in Vedic, Hittite, Greek, Latin
and Germanic.
SUPPLETION AND THE COPULA 5
by ἐγενόµην. As the pattern occurs in the Old IE languages of the Balkans, it
has to be concluded that it constitutes either a common innovation or a
convergent development (Kölligan 2007, 122).
*h1es- + *g̑ enh1- ‘to be born’: this is found only in Greek, cf. present 1sg εἰµί :
aorist ἐγενόµην and probably replaces the earlier suppletion with ἔπλετο.
*h1es- + *bʰu̯ eh2- ‘to grow, become, come into being’: this is the most
widespread suppletive pattern and can be found in Italic, e.g. Lat. present 1sg
sum : perfect fuī, Indo-Iranian, e.g. Old Indic present 1sg ásmi : aorist ábhuvam,
Balto-Slavic, e.g. Lith. present 1sg esù : past tense buvaũ, Old Church Slav.
present 1sg jesmь : aorist bychь, and Celtic, e.g. Old Irish present 3sg is :
preterite boí, Middle Welsh present 3sg yw/ys: preterite bu.
*h1es- + *h2u̯ es- ‘to stay, to pass the night’ + (*bʰu̯ eh2-) is restricted to Germanic,
cf. Gothic, present 1sg im : preterite was. The anlaut of Old High German
present b-im, b-is shows contamination with *bʰu̯ eh2- and suggests that *h2u̯ es-
replaced the latter as a suppletive root.
*h1es- + *nes- ‘to return, to come back’ + *steh2- ‘to stand’: 4 Tocharian
possesses a great variety of different roots, cf. Toch. B present 3sg ste, 3pl skente
(< *h1s-sk̑ e-to, *h1s-sk̑ o-nto) : perfect 1sg nesau : A, B subjunctive, preterite
tākā- (as if from *(s)teh2-k-). Also mäsk- can be used as a copula, cf. Toch. B
present 3sg mäsketär ‘is’ based on *mn̥ -ske/o- from PIE *men- ‘to stay, remain’,
in Lat. manēre and gr. µίµνω / µένω.
These suppletive patterns can be found in several language branches and even
more roots could be added if occurrences in individual languages and more
recent developments were to be considered. Their distribution suggests that the
copulas other than *h1es- emerged only after the PIE period within the individual
branches. Moreover, neighbouring languages tend to share the same patterns,
showing that areal factors clearly play a role for the selection of suppletive roots.
Does the distribution imply that *h1es-, which is the only copula in Anatolian,
was also the only copula in PIE? The evidence of the individual branches shows
that the copula experienced a considerable amount of formal and probably also
functional innovation. Some roots fell out of use and were replaced by others.
Thus the root(s) suppleting *h1es- in the proto-language might have gone lost.
*h1es- displays another peculiar feature: while all other roots are also attested as
full verbs, the only meaning traceable for *h1es- is ‘to be, exist’. Nevertheless,
this might not have been its original meaning.5 Some scholars argue that *h1es-
4
Adams (2013, 367-368), Pinault (1989, 133).
5
Shields (1992, 53-56), followed by Stassen (1997, 98-99), suggested that *h1es- originates from
the demonstrative/deictic stem *es-. In a sentence with copula deletion, in which *es- functioned as
6 BRITTA IRSLINGER
is identical with the root *h1eh1s- ‘to sit’, cf. Hitt. ēsa, Ved. ā ́ ste, Gr. ἧσται ‘sits’.
The unusual structure of the latter could be easily explained as originating from
reduplicated **h1e-h1s- (LIV 323). However, Kölligan (2007, 214) objects that
there are no cases in which a copula becomes a position verb, while the opposite
development is frequent. Furthermore, forms like HLuvian iša- ‘to sit’ point to
PIE *ē rather than *eh1. Kölligan suggests to reconstruct an original Narten
present *h1ḗs-ti / *h1és-n̥ ti ‘to sit; to be’ which split up into *h1ḗs-(n)toi̯ ‘sits,
seats oneself’ and *h1és-ti / *h1s-énti ‘is / are’. While the ē-grade forms
developed into a middle voice verb ‘to seat oneself’, the full-grade active forms
became a normal root present ‘to be’.
1.3 Suppletion: Sources and Developments
Suppletion of the copula is a common phenomenon not only in IE languages but
also cross-linguistically. The verbs entering into suppletive relations belong to
the group of “basic level concepts”, denoting basic human activities like HAVE,
GIVE, TAKE, HOLD, GET, STAND, GO, COME, BECOME, DO, MAKE.
They are conceptually simple as they are associated with a uniform mental
image and pragmatically salient as they correspond to elementary human
experiences. They often have multiple functions and may occur at the same time
as a full verb, as an auxiliary and as a copula. Due to their frequency and
semantic ambiguity, they are the starting point for various grammaticalisation
processes (Ströbel 2010, 162-3, Heine 1993, 27-37).
Regarding the function of suppletion, different explanations have been
suggested. Clancy (2010, 9-10) argues that verbs like BE or HAVE constitute
conceptual networks together with related notions. The different verbs
belonging to the network express the same concept in various contexts.
“The verbs and auxiliaries in the BECOMING-BEING-UNBECOMING network deal with
the concept of EXISTENCE with its specificities of TIME (verbs showing duration or
frequency) and SPACE (with concomitant notions of LOCATION and MOTION). Concepts
found in the network express the position, availability, and sensibility of objects, and they
establish relationships between various objects and the world and describe how those
relationships move and change.” (Clancy 2010, 10)
A given concept may thus absorb members of the network through the process
of suppletion, whereas the new lexical items may either replace older ones or
an anaphoric pronoun, it was reanalyzed as a verb. Although cross-linguistically copulas from 3sg
pronouns are frequent and this etymology could explain why *h1es- occurs only in the present tense,
it is problematic on the formal level. According to more up-to-date reconstructions, the
demonstrative stem *e-sm- (cf. Dunkel 2014, 196) and the verb *h1es- are not homonymous as
required by Shield’s argument.
SUPPLETION AND THE COPULA 7
find themselves a specific niche occurring in complementary distribution with
them (Clancy 2010, 23).
Ströbel (2010, 158-9, 178-187) claims that the replacement first occurs in face-
to-face communication, where the speaker tries to “visualize” a certain situation
with linguistic means. Here, the speaker may refer to the present situation, in
which the speaker, his addressee and their immediate surroundings are visible.
The body and its parts, like heads, arms, hands, legs, feet etc, are main points of
reference. However, if the speaker refers to situations in the past, in the future
or to common knowledge, he has to leave the situation on the linguistic level
and refer to “invisible” entities.
Reference to present situations predominantly induces the grammaticalisation
of full verbs into auxiliaries related to the “hands/arms” (have, hold, give, make,
do etc.) or to the “legs/feet” (stand, go come etc.). By metaphoric and
metonymic processes, they are used as more expressive equivalents of be,
linking the proposition to the situation more effectively. Once the semantic
content of a verb becomes bleached, it is replaced by another one in the order
indicated in table (3):
body → body part → body part + → body part +
(static) movement / action directed movement
“hands” HAVE MAKE GIVE
BE HOLD DO TAKE
“legs” STAND GO6 COME
STAY RETURN
Table 3: Tendencies of lexical verbs grammaticalising as copulas (Ströbel 2010, 185)
Ströbel’s model explains why suppletion is restricted to certain tenses and
supports the hypothesis that PIE *h1es- originally meant ‘to sit’.
2. The Habitual in Old English and Middle Welsh
Unlike other Germanic languages, Old English has two sets of copulas, known
as the e-forms and the b-forms (table 4). The paradigms of the Old English and
Middle Welsh copulas originate from the same PIE roots: OE is and MW ys/yw
are based on PIE *h1es-, while OE bið and MW. byð come from PIE *bhu̯ eh2-.
6
While in Ströbel’s model GO supplies BE, also the opposite development, i.e. BE supplying GO
is current, cf. Juge (2000, 191-193).
8 BRITTA IRSLINGER
Old English Middle Welsh Old Old High Old Gothic
Saxon German Norse
unmarked marked unmarked marked
biu bim em im
eom bīo wyf byðaf
eart bist wyt byðy bist bist est is
is bið ys, yw byð is(t) ist es ist
sint, bīoð ym byðwm sind(un) birum erom sijum
sind(on)
ywch byðwch birut eroþ sijuþ
ynt byðant sint ero sind
Table 4: Inflection of the verb ‘to be’, present indicative7
In Middle Welsh, ys/yw is the unmarked copula (COP-Y), used for concrete
situations (2). It contrasts with byð, which is the marked or habitual copula (COP-
B), denoting habitual or repeated actions (3) and also actions in the future (4)
(Schumacher 2009, 256-257).
concrete (2) R 1044.44, ms. 14th cent. (Schumacher 2009, 256)
situation: Stauell gyndylan ys tywyll heno.
unmarked hall [Cynddylan]GEN. COP-Y.PRS.3SG dark tonight
“Cynddylans hall, it is dark tonight.”
habitual: (3) proverb, R 1030.11 (Schumacher 2009, 256)
marked Bit amlwc marchawc.
COP-B.PRS.3SG clearly visible horseman
“A horseman is usually clearly visible.”
future: (4) CO 4.96-97, ms. c. 1350 (Schumacher 2009, 256)
marked ny byd gwaeth it yno nocet y Arthur yn y llys
NEG COP-B.3SG worse for.2SG there than for Arthur in ART court
“It will not be worse for you than for Arthur in the court.”
The same differentiations are assumed to underlie the Old English copulas, i.e.
the unmarked copula is (COP-E) denotes concrete situations (5), while the
marked bið (COP-B) is used for habitual or generic actions (6) or for those taking
place in the future (7) (Schumacher 2009, 257-8, Petré 2013, 306-8):
concrete (5) Beowulf 2532
situation: nis þæt ēower sīð
unmarked NEG-COP-E.PRS.3SG DEM.NS.N PRON.2PL.GS way.NS
“This is not your way.”
7
See Lühr (2015, 77-86) for a discussion of the Germanic forms.
SUPPLETION AND THE COPULA 9
habitual / (6) Beowulf 2541
generic: Ne bið swylc earges sīð
marked NEG COP-B.PRS.3SG such coward.GS way.NS
“Such is usually not the way of a coward.”
(7) Genesis 9:12
Đis bið ðæt tācn mīnes weddes
DEM COP-B.PRS.3SG DEM.NS.N sign.NS PRON.GS.1SG covenant.GS
“This will be the sign of my covenant.”
Given that both in Middle Welsh and in Old English copulas mark the same
semantic categories, contact influence seems possible. Furthermore, the habitual
is found in all Insular Celtic languages, but not in Germanic, which suggests that
Old English acquired it as a consequence of its contact with Celtic. While the
occurrence of this borrowing phenomenon is widely acknowledged, its dating
and the nature of the contact situation are controversial.
Schumacher (2009, 257) argues that it took place in the “Germano-Celtic
transition zone”, where neighbouring Celtic and Germanic peoples settled from
about 500 BC. Their common borders stretched for several hundred kilometres
through western and central Europe, from the mouth of the Rhine into the
western part of Slovakia. The intensive exchange between the two groups on the
cultural, technical and economic levels is evidenced by a large number of shared
lexemes, which are either mutual borrowings or common innovations. The
habitualis was transferred from Gaulish to West Germanic. Within the
continental dialects the paradigms merged, as evidenced by the b- in Old High
German bim etc., and the differentiation was lost again before the beginning of
the written tradition. Due to lack of evidence, it remains unclear however
whether the habitualis is already present in Gaulish or if it developed only later
in Insular Celtic.
In contrast to this, Lutz and Trudgill believe that the habitual/non-habitual
distinction emerged in Old English only after the Anglo-Saxons had reached
Britain. Trudgill (2011, 17) dates this process to the period from 420 to 600 AD,
when in many parts of Britain Old English speakers constituted a minority of
the population, outnumbered by Brittonic speakers. The feature was then
borrowed into Old English together with other structural features, such as the
grammaticalisation of the progressive aspect. Lutz (2010, 231ff.) considers it
improbable that intimate borrowing on roughly equal terms resulted in the
development of two full, functionally differentiated paradigms. The OE habitual
paradigm originated from substratal influence, i.e. it was created by English-
speaking Britons, who had acquired English imperfectly as adult L2 learners.
10 BRITTA IRSLINGER
The transfer to English occurred mainly through female house slaves in daily
contact with their masters and especially with their masters’ children.8
It is difficult to establish which of these two hypotheses is more credible. While
Lutz describes a more convincing sociolinguistic situation for contact-induced
borrowing, Schumacher (2009, 259) argues conclusively that the Old English b-
paradigm, in whatever function, dates back to Proto-West Germanic. Celtic
influence on Old English did not cause “additive complexification” but rather
prevented simplification.
This example illustrates a problem of contact hypotheses, especially between
cognate languages: as the morphological elements are present in both languages,
alternative explanations are possible.
3. Spanish, Portuguese and Catalan
3.1 Modern Languages
The paradigms of Spanish, Portuguese and Catalan ser result from the
syncretism of the VLat. verbs *ĕssĕre ‘to be’ and sedēre ‘to sit’, but the
etymologies of some forms remain doubtful. The infinitive ser ‘to be’ for
example, which is homonymous in all three languages, actually has different
origins. While Portuguese ser comes from V.-Lat. sedēre ‘to sit’, the Catalan
infinitive originates from VLat. *ĕssĕre ‘to be’. For Spanish ser, both VLat.
verbs have been suggested as a possible source, in addition to the hypothesis
that ser may come from sedēre with a strong influence of *ĕssĕre. 9 VLat.
*ĕssĕre and sedēre are based on the PIE roots *h1es- and *sed- ‘to sit’
respectively. Forms with f- can be found in tenses other than the present
indicative, cf. e.g. the preterite Span. fui, fuiste, fue, Port. fui, foste, foi, Cat. fui,
fores, fou etc. These paradigms are based on PIE *bʰu̯ eh2-, which is already
present in the Latin paradigm of esse. This latter root does however not play a
role in the hypotheses concerning the development of two-copula systems
mentioned above. Span., Portug. and Catal. estar, on the other hand, originated
from Lat. stāre ‘to stand’ from PIE *steh2-.
8
Petré (2013, 308), examining the distribution and merging of is and bið in Old and Middle English,
considers the question of a Celtic borrowing as irrelevant. However, he does acknowledge that the
distinction of the copulas was observed more strictly in the South, where the contact was much more
intense.
9
Span. ser < V-Lat. sedēre cf. e.g. Meyer-Lübke (1935, 259); < V-Lat. *ĕssĕre cf. e.g. Penny (2002,
191f.); < V-Lat. sedēre but strongly influenced by *ĕssĕre cf. e.g. Cano Aguilar (2002, 158).
SUPPLETION AND THE COPULA 11
Existentials are based on Lat. habēre ‘to have, to hold’, cf. Span. hay,10 Portug.
há, Catal. hi ha ‘there is, are’ (de Bruyne 1995, 583, Perini 2003, 278, Ramos
Alfajarín 2000, 29-30).
The function and distribution of ser and estar are similar in the three languages,
but not identical. Here I will only provide a brief description of the Spanish
system as given by de Bruyne (1995, 570-582) and Hengeveld (1986, 395-
400).11 Ser (COP-S) is obligatory in the following cases (ex. 8-13):
existence (8) De Bruyne (1995, 571)
Margarita si no se ríe, sonríe, feliz de ser.
Margarita if NEG REFL laugh.PRS.3SG smile.PRS.3SG happy to COP-S.INF
“If Margarita is not laughing, she is smiling, happy to be alive.”
identity (9) De Bruyne (1995, 570)
Vivir es sufrir.
live.INF COP-S.PRS.3SG suffer.INF
“Living is suffering.”
attribution (10) De Bruyne (1995, 570)
Soy abogado.
COP-S.PRS.1SG lawyer
“I am a lawyer.”
possession (11) De Bruyne (1995, 570)
Esta pluma es mía.
DEM.SG.F pen COP-S.PRS.3SG POSS.1SG.F
“This pen is mine.”
origin (12) De Bruyne (1995, 571)
Es de Madrid.
COP-S.PRS.3SG from Madrid
“He is from Madrid.”
material (13) De Bruyne (1995, 570)
La caja es de madera.
ART.SG.F box COP-S.PRS.3SG of wood
“The box is made of wood.”
time (14) Hengeveld (1986, 396)
La boda sera a las doce.
ART wedding COP-S.FUT.3SG at ART twelfe
“The wedding will be at twelve o'clock.”
Estar (COP-E), on the other hand, occurs in locative sentences, together with
locative noun phrases or adverbs as in (15, 16). An exception is the location of
10
From Lat. habet “has” + ibī “there” (Penny 2002, 191, 195).
11
See also Pustet (2003, 49-51), Maienborn 2005, Roby 2009, Dixon (2010, 175-6).
12 BRITTA IRSLINGER
an event, which requires the use of ser (17). In these cases, estar is still close to
its original meaning as a position verb.
position (15) De Bruyne (1995, 571)
with adverb Estoy aquí.
COP-E.PRS.1SG here
“I am here.”
position (16) De Bruyne (1995, 571)
with locative NP Carlos estaba en Segovia.
Carlos COP-E.IPF.3SG in Segovia
“Carlos was in Segovia.”
exception: (17) Hengeveld (1986, 397)
place of event12 La reunión es en la sala catorce.
ART meeting COP-S.PRS.3SG in ART room fourteen
“The meeting is in room 14.”
With predicative adjectives, ser denotes a permanent quality or an essential
attribute (18) as opposed to estar, which denotes behaviour or a temporary
attribute (19). Only in this case estar can occur exactly in the same context as
ser. Following Carlson 1977, temporary attributes are also called “stage level
predicates”, permanent ones “individual level predicates”.
behaviour, temporary attribute (18) De Bruyne (1995, 575)
estar + adjective Algunos días estaba alegre.
some.PL day.PL COP2.IPF.3SG happy
“Some days he was happy.”
essential characteristic, (19) De Bruyne (1995, 574)
permanent attribute Bonifacio era muy allegre.
ser + adjective Bonifacio COP1.IPF.3SG very happy
“Bonifacio was very happy.” (“B. was a happy person.”)
The distinction permanent vs. temporary is also present in some of the previous
examples. For example, the origin of a referent or the material something is
made of are permanent attributes, while being in a place is often temporary.
It is possible to express a temporary attribute with estar, but in this case the
copula complement is introduced by the preposition de 'of, from' (ex. 20):
temporary attribute (20) Stassen (2005, 483)
estar + de + NP Julia está de enfermera (en Madrid).
Julia COP-E.PRS.3SG PREP nurse in Madrid
“Julia is a nurse (in Madrid).” / “Julia works as a nurse.”
12
See the discussion of this exception in Roby (2009, 75-76).
SUPPLETION AND THE COPULA 13
In Portuguese, no preposition is necessary, i.e. estar is used with non-locative
NPs to denote a temporary attribute or occupation (21), while ser denotes a
permanent attribute or occupation (22). The aspectual distinction thus spread
from adjectival predicates to nominal ones. The feature can also be found in the
Spanish dialect of Cabrales (Asturia) with complements like braceru 'farm
hands, farmworker' denoting temporary occupations (Hengeveld 1991, 88,
Mateus / Brito / Duarte / Faria 1983, 138).
temporary attribute (21) Hengeveld (1991, 88)
estar + NP O Luís está professor em Beja.
ART.SG.M Luís COP-E.PRS.3SG professor in Beja.
“Luís is a professor in Beja.” / “L. works as a professor.”
permanent attribute (22) Hengeveld (1991, 88)
ser + NP O Luís é professor em Beja.
ART.SG.M Luís COP-S.PRS.3SG professor in Beja.
“Luís is professor in Beja.”
Cleft sentences are always introduced by ser (23), while the progressive is
formed with estar + gerund (24). This use of the position verb in the progressive
has parallels in Italian, Basque and Celtic.13
contrast focus (23) De Bruyne (1995, 570)
cleft sentence Es Juan que/quien habla.
COP-S.PRS.3SG Juan REL speak.PRS.3SG
“It’s JUAN speaking.”
progressive (24) De Bruyne (1995, 571)
estar + gerund Pedro está trabajando.
Pedro COP-E.PRS.3SG work.GERUND
“Pedro is working.”
3.2 Old Spanish
This distribution of ser and estar is relatively recent. In Old Spanish and
occasionally until the end of the 17th century, ser occurs with locative
complements (25) besides estar (26). In ex. (27) and (28), ser is used to denote
temporary attributes, as the use of estar in this function was still very rare in Old
Spanish. The two-copula system of Modern Spanish is thus a recent
development (Pountain 1985, 350-353, Vañó-Cerdá 1982, 246, Remberger /
González-Vilbazo 2007, 207).14
13
On periphrastic constructions with ser or estar + participle see Hengeveld 1986.
14
See also Bouzet (1953, 37-58), Lapesa (2008, 189, 338). For Old Catalan see Ramos Alfajarín
(2000, 135), for Old Portuguese see Pountain (1982, 153).
14 BRITTA IRSLINGER
Position (25) Cantar de Mio Cid, V. 1094, c. 1200 (Lapesa 2008, 189)
el Señor que es en çielo
ART lord REL COP-S.PRS.3SG in heaven
“the lord who is in heaven”
Position (26) Cantar de Mio Cid, V. 330, c. 1200 (Lapesa 2008, 189)
Padre que en cielo estás
father REL in heaven COP-E.PRS.2SG
“father, (you) who are in heaven”
attribution: (27) Cantar de Mio Cid, V. 1924, c. 1200
temporary Es pagado, e da vos su amor
COP-S.PRS.3SG content and give.PRS.3SG PRON.2PL POSS.3SG.M love
“He is contend and he gives you his love”
(28) Poema de Fernán Gonzáles 136c, 13th cent. (Vañó-Cerdá 1982, 246)
Folgaron e dormieron que eran muy cansados
rest.PRT.3PL and sleep.PRT.3PL as COP-S.IPF.3PL very tired
“They rested and slept because they were very tired”
4. Basque
In Basque two different systems are in use. The Eastern dialects, spoken in
France, have a one-copula system: the copula izan ‘to be’15 (COP-I) is used in all
contexts, cf. (29) with a locative complement (Trask 1997, 292).
position (29) Trask (1997, 292)
Aita etxean d-a.
father house.LOC 3SG.ABS-COP-I.PRS
“Father is in the house.”
The Western dialects, spoken in Spain, employ egon ‘to be in a certain place,
state or condition’ (COP-E) as a second copula that shows a distribution similar
to that of Spanish estar. However, egon is not strictly speaking the verb STAND,
as it means ‘to wait, stay, remain’ in the Eastern dialects and it still has this
meaning when used as an imperative in the West. The etymologies of the Basque
verbs are unknown (Trask 2008, 235, 165, de Rijk 2008, 116).
Egon denotes a position (30), but also temporary states and qualities when used
with a locative NP (31) or with predicative adjectives (32). Like in Portuguese,
egon can also be used with nouns to denote a temporary state (33). Further, egon
occurs as an auxiliary to form the progressive (34).
15
The present tense forms of this verb are highly irregular, probably due to stem suppletion (Trask
1997, 232).
SUPPLETION AND THE COPULA 15
position (30) Zabala (2010, 426)
Amaia zer egin ez dakiela d-ago.
Amaia what do NEG know.COMP 3SG.ABS-COP-E.PRS
“Amaia just stands/is there not knowing what to do.”
temporary (31) Zabala (2010, 427)
attribute: Beñat etxean / lanean / langabezian d-ago.
with locative NP Beñat home.LOC / work.LOC / unemployment.LOC 3SG.ABS-COP-E.PRS
“Beñat is at home / working / unemployed.”
with adjective (32) Zabala (2010, 427)
Mikel urduri d-ago.
Mikel nervous 3SG.ABS-COP-E.PRS
“Mikel is nervous.”
with NP (33) Zabala (2010, 427)
Jon irakasle d-ago MITn.
Jon lecturer COP-E.PRS.3 SG MIT.LOC
“Jon is (working as) a lecturer at MIT.”
progressive (34) Zabala (2010, 427)
Nerea abesten d-ago.
Nerea sing.NOM.LOC 3SG.ABS-COP-E.PRS
“Nerea is singing.”
Izan is used for permanent attributes (35), but also to denote the location of an
event like in Spanish (36). Additionally, about twenty current nouns denoting
temporary mental or physical states are constructed with izan in all dialects (37),
cf. e.g. beldur ‘fear’ : beldur izan ‘to be afraid’, eri ‘illness’ : eri izan ‘to be ill’.
Further examples are haserre ‘anger’, gose ‘hunger’ or egarri ‘thirst’ (de Rijk
2008, 154-55). This group constitutes an important exception from the aspectual
differentiations of the copulas.
permanent attribute (35) Zabala (2010, 427)
Miren bilbotarra d-a.
Miren Bilbaoan.DET 3SG.ABS-COP-I.PRS
“Miren is from Bilbao (is a Bilbaoan).”
place of event (36) Zabala (2010, 429)
Jaia Mirenen etxean d-a.
party.DET Miren.GEN house.LOC 3SG.ABS-COP-I.PRS
“The party is at Miren’s house.”
certain temporary mental or (37) Zabala (2010, 428)
physical states Haurra beldur d-a.
child.DET fear 3SG.ABS-COP-I.PRS
“The child is frightened.”
16 BRITTA IRSLINGER
The origin of the Western Basque system is controversial, because the lack of
medieval evidence makes it impossible to date it. While Trask (1997, 292-293)
and others think that it emerged due to Spanish influence, Vennemann (2010,
393) claims that it was already present in Proto-Basque, but was lost in the
Eastern dialects due to contact with French or Gascon.
However, in view of the distribution and diachronic spread of estar this latter
hypothesis is not convincing. According to Hengeveld (1992: 245), estar spread
from locative predication to the adjectival and then to the nominal one:
Locative > Adjective > Nominal (> Possessive)
Different varieties of Ibero-Romance display different stages of this
development (table 4). For example, estar is still restricted to locative predicates
in Judeo-Spanish, but can be used with adjectivals in Spanish, Catalan and
Galician. Only in Portuguese, estar spread to nominal predicates. The next step,
which has however not yet occurred in any of the languages, would be the spread
to possessive predicates.
Adjectivals Nominals
Judeo-Spanish – –
Catalan + –
Spanish + –
Galician + –
Portuguese + +
Western Basque + +
Table 4: Distribution of copula verbs derived from position verbs in Ibero-Romance and
Basque (adapted from Hengeveld 1992, 245)
If the Western Basque system were the one which influenced Ibero-Romance,
one would expect to find estar with nominal predicates in the neighbouring
languages or varieties and at an earlier date. On the contrary, the data point to a
gradual spread of estar in Ibero-Romance and, as far as the origin of the Western
Basque system is concerned, to a possible contact-induced replication with
overgeneralization16 of the ser : estar distinction.
5. Irish
5.1 Modern Irish
Irish grammar traditionally distinguishes between the “copula” and the
“substantive verb”. The proper copula is proclitic, precedes the copula
16
Cf. for this process Heine / Kuteva (2005, 116).
SUPPLETION AND THE COPULA 17
complement and may be dropped in positive clauses in the present tense. Its
inflected forms of the indicative present are based on PIE *h1es-. The other
forms are supplied by PIE *bʰu̯ eh2- and thus have initial b-. While in Old Irish
the copula was still inflected, it is uninflected and defective in Modern Irish,
with is used for present tense and ba for past and conditional. In negative,
interrogative and subordinate clauses the copula fuses with the respective
preverbal particles and conjunctions, which are mostly not analysable already in
Old Irish, e.g. OIr. ní (ni) ‘is not’ < *nītes < *ne eti esti (GOI 483-494,
Schumacher 2004, 311).
Diachronically, the copula continuously developed from an inflected verb into
a functional morpheme, to the point that in Modern Irish its “verbal” features
have almost disappeared.
The substantive verb, on the other hand, is stressed and possesses also in Modern
Irish a full paradigm and a residue of lexical semantics. Modern Irish tá,
originating from the Old Irish compound verb attá < *ad-tá ‘to be present, to
stand by’ is based on PCl. *tā-i̯ e/o- from PIE *steh2-. Its paradigm is highly
suppletive, as tá is restricted to positive main clauses in the present indicative.
In negative, interrogative or subordinate clauses, the so-called dependent form
OIr. fil, ModIr. bhfuil is used. OIr. fil was originally the imperative ‘see!’ of a
verb based on PIE root *u̯ el- ‘to see’, and in archaic texts can also be found to
indicate existence. All other tenses and moods are supplied by forms based on
*bʰu̯ eh2-. In Old Irish, a form of gaibid ‘to take’ (PIE *gʰeHb- ‘to take’), i.e. the
perfect ron(d)-gab ‘has taken it’ with an infixed 3sg neuter pronoun occurs in
nasalizing relative clauses as a copula (GOI 476-483, Schumacher 2004, 623-
26).
Contrary to Spanish, Portuguese and Catalan, OIr. attá does not express ‘to
stand’ as a bodily posture.
In Modern Irish, is (COP-I) expresses identity (38) and attribution (39) with
nominal and pronominal complements. In both cases, permanent attributes are
involved. Is also introduces cleft sentences (40), which are frequent in all periods
of Irish (Ó Siadhail 1989, 219-252, Doyle 2001, 65-67, Wigger 2003, 264-64,
Nolan 2012, 204-213, NIG 122-125, GG 167-170, 175-192).
identity (38) Ó Siadhail (1989, 227)
Is mé an múinteoir.
COP-I.PRS PRON.1SG ART teacher
“I am the teacher.”
18 BRITTA IRSLINGER
classification (39) Ó Siadhail (1989, 220)
/ attribution Is bádóir é.
COP-I.PRS boatman PRON.3SG.m
“He is a boatman.”
contrast focus (40) Ó Siadhail (1989, 237)
cleft sentence Is tinn atá sé.
COP-I.PRS sick COP-T.PRS.REL PRON.3SG.M
“He is SICK.” (lit. “It is sick that he is.”)
Tá (COP-T) denotes existence (41) or a position (42) expressed by a locative NP
or adverb. Tá can also express temporary states and attributes or the notion ‘to
have become’. However, this case requires the use of a locative complement
consisting of in ‘in’ + possessive + either a verbal noun (43) or a noun (44). The
latter construction is mainly restricted to denotations for roles or professions,
but it is also possible to find exceptions like (45) or more idiomatic expressions
like Tá sé ina lá ‘It is day’ (lit. ‘It is in its day’.). Finally, tá + ag ‘at’ + verbal
noun is used to form the progressive. If the progressive phrase is in the past tense
(46), a form based on *bʰu̯ eh2- is used.
existence (41) NIG 117
• tá + locative NP or adverb Tá trí phersa i nDia.
COP-T.PRS three person.SG in god.SG
“There are three persons in god.”
position (42) NIG 117
• tá + locative NP or adverb Tá Seán ar scoil. / anseo
COP-T.PRS Seán on school / here
“Seán is at school / here.”
temporary state, process (43) NIG 117
• tá + locative NP Tá Seán ina choladh.
(i ‘in’ + possessive COP-T.PRS Seán in in-POSS.3SG.M sleeping.VN
+ verbal noun) “Seán is sleeping.” (lit.: “Seán is in his sleeping.”)
• tá + locative NP (44) NIG 117
(i ‘in’ + possessive + noun) Tá Máire ina múinteoir anois.
COP-T.PRS Máire in-POSS.3SG.F teacher now
“Máire is a teacher now.”/ “M. has become a teacher.”
(lit. “Máire is in her teacher now.”)
permanent state (45) NIG 118
• tá + locative NP Tá an tAthair ina Dhia.
(i ‘in’ + possessive + noun) COP-T.PRS ART father in-POSS.3SG.M god
“The Father is God.”
SUPPLETION AND THE COPULA 19
progressive (46) NIG 93
tá + ag ‘at’ + verbal noun Bhí Peadar ag rith.
COP-T.PRT Peadar at run.VN
“Peadar was running.”
Is or tá occur in a number of other constructions in which their distribution is
not predominantly governed by the permanent vs temporary distinction.
Discourse-pragmatic features play a role with regard to adjective complements.
While tá introduces a pragmatically neutral clause with the adjective in final
position (47), the construction with is requires that the adjective be fronted and
thus focussed (48). In this exclamatory construction expressing a subjective
view, only adjectives denoting permanent qualities are permitted.
pragmatically neutral (47) Ó Siadhail (1989, 236)
tá + subject + adjective Tá sé tinn.
COP-T.PRS PRON.3SG.M sick
“He is sick.”
exclamatory (48) Ó Siadhail (1989, 229)
is + adjective + subject Is maith é!
COP-I.PRS good PRON.3SG.M
“He is (so) good!”
Comparative and superlative constructions only allow the use of is (49). The
same is true for impersonal constructions like is féidir ‘it is possible’ or is
deacair ‘it is difficult’ as well as for modal constructions like is maith liom ‘I
like’ or is fearr liom ‘I prefer’ (50).17 Additionally, a small number of adjectives
(aisteach ‘odd’, beag ‘small’, cosúil ‘similar’, fíor ‘true’, fiú ‘worthwhile’, fuar
‘cold’, greannmhar ‘funny’, ionann ‘equivalent’, leor ‘sufficient’, mall ‘slow,
sluggish’, maith ‘good’, mór ‘big’ and olc ‘evil’) can still be constructed with
the copula (51). This construction of adjectives denoting permanent qualities
constitutes a relic of Old Irish, where the substantive verb could not be combined
with adjectives. For the same reason, certain frequent adjectives like maith
‘good’, dona ‘bad’ or deas ‘nice’ occur with the adverbial marker go, when used
as complements of tá (52) (Ó Siadhail 1989, 230-32, Doherty 1996, 32, Dillon,
1927, 317).
comparative and (49) Wigger (2003, 264)
superlative clauses Is measa Peadar ná Pól.
COP-I.PRS worse Peadar than Pól
“Peadar is worse than Pól.”
17
While the use of certain adjectives with is or tá is to some degree idiomatic, GG 182 points out
that is mostly denotes more permanent qualities and states and tá more temporary ones.
20 BRITTA IRSLINGER
modal constructions (50) Wigger (2003, 264)
Ba mhaith liom toitín a chaitheamh.
COP-I.KOND good with.1SG cigarette to consume.VN
“I would like to smoke a cigarette.”
various adjectives (51) Doherty (1996, 37)
denoting permanent Is aisteach agus is iontach bealaigh Dé.
qualities (relics) COP-I.PRS strange and COP-I.PRS wonderful ways.PL god.GS
“The ways of God are strange and wonderful.”
with adverbial marker (52) Wigger (2003, 264)
go Tá sí go deas.
COP-T.PRS.3SG PRON.3SG.F ADV nice
“She is nice.”
Some scholars have claimed that the use of is : tá with adjectival predicates in
Irish corresponds to the use of ser : estar in Spanish. According to Greene (1966,
41-42), cited by Stassen (1997, 180-1), examples like (47) refer to a temporary
state of sickness, while (52) indicates that somebody is not necessarily nice in
general by is being nice in a given moment. Constructed with is, the same
adjectives would express inherent qualities. 18 He admits that “it is doubtful
whether many speakers feel a distinction between them nowadays.”
However, one wonders if this distinction, which could be exemplified by a small
number of minimal pairs, was ever productive. Ex. (53) from the grammar of
the Christian Brothers published around 1920 should not be possible if Greene
was right. The adjectives constructed with the copula listed above constitute a
closed set. Thus all other adjectives denoting permanent attributes have to
appear with the substantive verb, i.e. the speaker is not free to choose between
is and tá to express an aspectual distinction (Doherty 1996, 37).
permanent attribute (53) Graiméar na Gaedhilge (s.a., about 1920, 205)
Tá an bhó mhór dubh.
COP-T.PRS.3SG ART cow big black.
“The big cow is black.”
There are two semantically different constructions to denote possession and
ownership respectively. Tá + ag ‘at’ means ‘to have, to possess’ (11), while is
+ le ‘with’ denotes ownership (12). Nolan (2012, 210) calls the latter function
“ownership identity”. According to the Dictionary of the Irish language (eDIL
O 85b), possession is correlated with the feature “temporary”. This claim has to
18
See also Hickey (1968, 222-224) and Devitt (1990, 110), who gives the following ex. without
citing a source: Tá an páipéar bán. “The paper is white.” (i.e. blank; not written on) vs. Is bán an
páipéar é. “The paper is white.” (its whiteness is inherent). According to Ó Siadhail (1989, 229),
the latter sentence is exclamatory, i.e. pragmatically marked.
SUPPLETION AND THE COPULA 21
be questioned, as the concepts of possession and ownership do not seem to be
intrinsically linked with the features temporary and permanent (see also the Old
Irish examples 67-69 below).
possession (54) Ó Siadhail (1989, 233)
tá + ag ‘at’ Tá carr ag Cáit.
COP-T.PRS car at Cáit
“Cáit has/owns a car.” (lit. “There is a car at Cáit.”)
ownership (55) Ó Siadhail (1989, 233)
is + le ‘with’ + Is le Cáit an carr.
definite noun COP-I.PRS with Cáit ART car
“The car is Cáit’s / Cáit owns the car.” (lit. “The car is with Cáit.”)
5.2 Old Irish
With regard to nominal complements the use of is and attá in Old Irish largely
corresponds to the Modern Irish one. Is expresses identity as in (56) and (57). In
the latter example, the copula has been dropped. Additionally, is is used for
attribution (58) and to introduce cleft sentences (59) (GOI 475-494, McCone
2005, 39-51, Dillon 1927, 1928).
identity (56) Wb. 1a2, 8th cent.
is Dia-som dom-sa
COP-I.PRS.3SG god.NS-EMPH to.1SG-EMPH
“He is god to me.”
identity (57) TBDD, 1.1-2, ms. c. 1100
copula dropped Echaid Feidlech a ainm.
Echaid.NS Feidlech.NS POSS.3SG.M name.NS
“Echaid Feidlech (was) his name.”
classification (58) Wb. 27c22, 8th cent.
/ attribution am cimbid-se
COP-I.PRS.1SG captive.NS-EMPH
“I am a captive”
contrast focus (59) Wb. 27c22, 8th cent.
cleft sentence is airi am cimbid-se
COP-I.PRS.3SG therefore COP-I.PRS.1SG captive.NS-EMPH
“It is therefore I am a captive”
OIr. attá denotes existence, cf. ex. (60) in present tense and (61) in preterite with
a form based on *bʰu̯ eh2-. With a locative NP or adverb, it denotes position, cf.
ex. (62) with suppletive fil. Old Irish also displays a construction with locative
NP, attá + i ‘in’ + possessive + verbal noun (63) that expresses a temporary
state. On the other hand, attá + i ‘in’ + possessive + noun does not appear until
22 BRITTA IRSLINGER
1100. Tá sé i n-a righ means “He was king (at the moment)” or “He had become
king” (64) (Dillon 1928, 309). The progressive (65) can be already found in Old
Irish, although it is not yet as frequent.
existence (60) Wb. 2c21, 8th cent.
a:taat da n-orpe
COP-T.PRS.3PL two heritage.NS
“There are two heritages”
(61) TBDD, 1.1-2, ms. c. 1100
boí rí amrae airegdae for Érinn
COP-T.PRT.3SG. king.NS wonderful.NS.Mnoble.NS.M over Ireland.AS
“There was (once) a wonderful noble king over Ireland.”
position (62) TBC I 16.503, ms. c. 1100
attá + locative in:fil mo phopa Conchobor isind
NP or adverb PT.-COP-T.DEP.PRS.3SG POSS.1SG daddy Conchobor in-the
ármaig se?
battle-field EMPH
“Is my daddy Conchobor in this battle field?”
temporary state (63) Wb. 13a12, 8th cent.
attá + locative bíis inna suidiu
NP COP-T.HAB.PRS.3SG. in -POSS.3SG.M sit.VN
“he uses to be sitting”
(64) PH 2572, c. 1150, Middle Irish
robói ... 'n-a rig
COP-T.PRT.3SG. in-POSS.3SG.M king.NS
“He was king (at the moment).” / “He has become a king.”
progressive (65) TBC I 14.437, ms. c. 1100
attá + oc ‘at’ + bámmar oc imbirt fidchille
verbal noun COP-T.PRT.1PL at play.VN fidchell.GS
“We were playing fidchell (a board game)”
Possession can be expressed in different ways. The possessor can be coded as a
pronoun suffixed to *táith, the otherwise unattested 3sg absolute form of the
present indicative of attá (66) (GOI 271). Additionally, there are the
constructions attá + oc to denote possession (67) and is + le to denote ownership
(69) like in ModIr. Ex. (68) seems to use a combination of the two and denotes
possession, while elsewhere attá + le expresses company. Again, there seems to
be no correlation with the features “permanent” and “temporary”.
*táith + suffixed (66) Thes. ii 293.19
pronoun táith-iunn
COP-T.PRS.ABS.3SG-PRON.1PL
“we have, there is to us”
SUPPLETION AND THE COPULA 23
• tá + oc ‘at’ (67) TBC II 862
árchú maith fil ocom
war-hound good COP-T.PRS.3SG.REL at.1SG.
“a good war-hound that I have”
• tá + le ‘with’ (68) TBC I 572, ms. c. 1100
atá árchú lemsa
COP-T.PRS.3SG war-hound with.1SG-EMPH
“I have a war-hound” (lit.: “There is a war-hound with me”)
• is + le ‘with’ (69) LU 8974; FB 73, ms. c. 1100
is lim-sa in cauradmir
COP-I.PRS.3SG with.1SG-EMPH ART champion’s portion.NS
“The champion’s portion belongs to me.”
Contrary to Modern Irish, is has to be used with adjective complements, cf. ex.
(70), which denotes a temporary state. Using attá would not be possible here.
attribution (70) Sg. p. 112, Thes. ii 290.4, 9th cent.
adjective Is acher in gáith innocht
COP-I.PRS.3SG bitter.NS ART wind.NS tonight
“Bitter is the wind tonight.”
There are however several deviations from these rules. In the 7th-century
Auraicept na nÉces, the finite forms of Latin esse are translated using the
substantive verb and not the copula (Dillon 1928, 329-332).
What is more problematic is the apparent use of the substantive verb with NPs
and adjectives in the 7th-century Amra Choluim Chille and in several other Old
Irish texts. Dillon (1928, 337) suggests to explain boe and boi in ex. (71) and
(72) not as instances of the substantive verb but as variants of the copula ba,
which actually occurs in the same text in identical constructions. They could be
interpreted as deliberate archaisms reflecting the unreduced form of the copula.
attribution (71) ACC 26, 7th cent.
NP boe sab suithe cech dind
COP-T.PRT.3SG prince knowledge.GS each high place
“he was the prince of knowledge (in) every high place.”
attribution (72) ACC 74, 75, 7th cent.
adjective boi cath, boi cast
COP-T.PRT.3SG wise COP-T.PRT.3SG chaste
“he was wise, he was chaste”
This survey showed that the distribution of Irish is and tá only partly
corresponds to the Ibero-Romance distinction between ser and estar. As
expected, nominal identificational and attributive copula complements occur
with is, while attá and tá are found with locative ones and in the progressive.
24 BRITTA IRSLINGER
However, the use of tá spread to adjective copula complements only very
recently, whereas at the same time both copulas are found in possessive
constructions already in Old Irish. This contradicts the predicate hierarchy for
attributive predicates established by Hengeveld (1992, 130). This hierarchy
implies that if a language can use possessive predicates predicatively, it can also
do so with any predicate on the left of the scale.
Locative > Adjective > Nominal > Possessive
For is this prediction is valid, given that it has been replaced by attá with locative
predicates. On the other hand, attá seems to have spread differently than its
Ibero-Romance equivalent, as the new copula is found with nominals and
possessives, but has skipped the adjectivals.
It should be noted, however, that the constructions including nominals and
possessives are not identical to the Ibero-Romance ones. The types Tá tú i do
choladh and Tá mé i mo mhúnteoir contain locative constructions and are thus
to be considered as the expansion of the positional function. The possessive
construction of the type Tá teach ag an mbean might be related to the existential
function, cf. the English literal translation “There is a house at the woman.” This
type is not contained in Hengeveld’s predicate hierarchy, which refers to non-
presentative, i.e. non-existential, uses. Due to the lack of evidence, it remains
unclear whether the existential function spread to the possessive one and may
thus be considered as a short-cut on the predicate hierarchy. If one disregards
the existential and similar constructions (+E), Irish tá is still mainly restricted to
locative constructions.
Existence Position Adjectivals Nominals Possessive
Old Irish is – – + + +
attá + + – –/(+L) –/(+E)
Mod. Irish is – – –/+ + +
tá + + –/+ –/(+L) –/(+E)
Table 5: Predication in Irish
Nevertheless, constructions that resemble the existential type and consist of tá
+ noun + inflected preposition coding the agent are extremely frequent. As they
code concepts that are expressed verbally in other languages, they are
comparable to light verb constructions. Some of them are mere stylistic
variations to simple verbs, but in many cases, especially with regard to mental
and physical actions and states, no equivalent verbs are available (73). They
outnumber constructions with is (74) (Wigger 2003, 258).
SUPPLETION AND THE COPULA 25
(73) tá + N + ag ‘at’ tá a fhios agam ‘I know’
tá súil agam ‘I hope’
tá an ghráin agam ‘I hate’
tá + N + ar ‘on’ tá fearg orm ‘I’m angry’
tá ocras orm ‘I’m hungry’
tá aiféal orm ‘I regret’
(74) is + A + le ‘with’ is maith liom ‘I like’
is fearr liom ‘I prefer’
6. Old French
For the advocates of the Vasconic hypothesis a decisive piece of evidence comes
from Old French. The Vulgar Latin verbs essere and stāre developed into the
very similar Old French verbs estre ‘to be’ < *essre and ester ‘to stand’ < *estar.
While OFr. estre became ModFr. être, OFr. ester is only preserved in phrases
like ester en jugement ‘to be on trial, to go to court’, cf. Germ. vor Gericht stehen
(lit. ‘to stand before court’).
From the 13th century onwards, forms based on the stem est- replaced the
inherited imperfect. Their origin is controversial. Many scholars derive them
from ester, which also supplied the present and past participles OFr. estant
‘being’ (ModFr. étant) and OFr. estet ‘been’ (ModFr. été), but not all forms
show the expected outcome. As Lat. stabat is preserved as esteve in some
dialects, the 3sg estoit might be formed analogically from the infinitive est-
+ -oit on the model of vendre : vendoit (table 6, Revol 2000, 137-8, Bonnard /
Regnier 1989, 110-1). Pountain (1982, 147) rejects the idea that ester supplies
the new imperfect of estre.
Latin Old French imperfect etymology
inherited form new form
1S eram > (i)ere → esteie > estoie < *stebam < Lat. stabam?
2S eras > (i)eres → esteies
3S erat > (i)ere, (i)ert → esteit, estoit < Lat. stabat or est- + -oie?
1P eramus > erïens/erïons → estïens/estïons
2P eratis > erïez → estïez
3P erant > (i)erent → esteint
PPA estant < Lat. stantem/stando
PPP estet < Lat. statum
Table 6: Old and new forms of the Old French imperfect
Vennemann (2010, 393-4) and Trudgill (2011, 12-13) support the suppletion
hypothesis and claim that suppletion was only possible because, on the one hand,
both verbs had become very similar on the formal level due to phonological
changes and, on the other, their semantics had converged. The latter
26 BRITTA IRSLINGER
development implies that ester ‘to stand’ was semantically bleached and thus
could take over the function of a copula. However, before the fusion occurred,
there must have been an aspectual distinction between the two. Vennemann
supposes that Old French distinguished between the copulas “in a similar way
to other Romance languages”, while Trudgill argues that the difference was
between habitual (corresponding to permanent, cf. Trudgill 2010, 4) and non-
habitual (corresponding to temporary). Trudgill also speculates that if this
distinction was present in Celtic (and perhaps ultimately in Vasconic) and
entered North-western Romance, it “certainly would represent a good example
of complexification vis-à-vis earlier forms of Latin”. These far reaching
conclusions are however not necessarily obvious.
Bonnard / Reigner (1989, 111) claim that estre and ester were semantically
identical and could be used in the same contexts, like in ex. (75), where a form
of estre ‘to be’ occurs with the locative complement d’une part ‘on one side’
and both estre and ester are used with predicative adjectives.
(75) Queste 246, 9, c. 1220
Car d’ une part ert la forest qui granz estoit et desvoiable;
For on one side estre.IPF.3SG ART forest.SG REL big ester.IPF.3SG and impassable
et autre part avoit deus roches qui estoient hautes
and other side have.IPF.3SG two rock.PL REL ester.IPF.3PL high.PL
et anciennes,et d’ autre part l’ eve qui ert parfonde et noire.
and ancient.PL and on other side ART water rel estre.IPF.3SG deep and black
“Because on one side, there was the forest which was big and impassable, and on the other
side, there were two rocks which were high and ancient, and on the other side, there was the
water, which was deep and black.”
In opposition to this view, Pountain (1982, 146-7) and Stengaard (1991, 320-
329) established the following semantics of ester: 1. ‘to stand’, 2. ‘to stay’, 3.
‘to stop’, 4. ‘to live’ 5. reflexive ‘to stand still’. Stengaard (1991, 320) claims
that this verb also conveys the meaning ‘to stand up’, which is actually its main
meaning. It is striking that in ex. (75) forest ‘forest’ and roches ‘rocks’, i.e. high
objects in a vertical position, are the subjects of ester, while the flat entity eve
‘water’ combines with estre. Because of this, rather than evidencing a merger,
ex. (75) could attest the semantic difference between the two verbs.
Although the suppletion took place only in the imperfect, Tobler-Lommatzsch
III 1382-3 gives a handful of examples in other tenses where ester could be
interpreted as a copula. In ex. (76), ester occurs with a nominal attribute,
whereas another version of the text has the corresponding form of estre ‘to be’.
If estommes meant ‘we are’, it would occur with a nominal complement
indicating a permanent attribute. One should however consider the possibility of
SUPPLETION AND THE COPULA 27
ester being used with its original meaning, i.e. ‘stand firmly, remain steadfast as
a wise cleric’.
(76) Barl. u. Jos. 6392-93, 1200-1250 (Tobler-Lommatzsch III 1382-3)
Et nous, ki sage clerc estommes / variant: letré somes
and PRON.1PL REL wise cleric.SG ester.PRS.1PL / learned estre.PRS.1PL
A autres dex ne savons tendre.
to other.PL god.PL NEG know.PRS.1PL turn.INF
“And we who are / stand (as a) wise cleric, we are not able to turn to other gods.”
Pountain’s (1982, 147) argues convincingly that ester did not achieve auxiliary
or copular status in Old French and therefore disappeared.
7. Vulgar Latin
If the hypothesis of Celtic influence on Gallo-Romance were right, this influence
should of course have manifested itself much earlier than in Old French, as
Gaulish is believed to have died out by the end of the 6th century AD at the latest.
In Vulgar Latin texts ranging from the 3rd to the 8th century, the use of stāre in
contexts typical for copulas increases, both in the Latin of Gaul and in other
varieties. Nevertheless, stāre cannot be considered as a copula, but is attested in
the following meanings: 1. ‘to stand upright’, 2. ‘to remain (standing)’, 3. ‘to
stand in a place’, 4. ‘to be in a state’ (Remberger / González-Vilbazo 2007, 204-
206, Stengaard 1991, 29-59, 57).
The author of ex. (77), Venantius Fortunatus (c. 530-609 AD), was born in
Northern Italy and became the court poet of the bishop of Poitiers. His language
might thus reflect the Gallo-Latin usage of the 6th century. In the example, stāre
expresses identity like a copula. However, it is unlikely that god, the subject of
the clause, would be referred to as being the temporary originator and creator of
men. Stāre rather seems to indicate immobility, cf. Stengaard (1991, 59).19
(77) Venantius Fortunatus, Carm. V, 5, 53; 6th cent.
qui tuus, ipse meus stat conditor atque creator
REL.NS POSS.2SG self.NS.M POSS.1SG stand.PRS.3SG originator.NS and creator.NS
“who is/stands as the originator and creator of yours and also of mine”
8. Brittonic
This section will trace the fate of *steh2- in the Brittonic languages, as well as
the rise of a new locative copula, which so far has not been considered in the
present context.
19
Fels (2006, 119) translates: “Dieser dein Urheber und Schöpfer zeigt sich zugleich auch als
meiner” with German sich zeigen ‘to appear, become apparent’.
28 BRITTA IRSLINGER
8.1 Old British
A two-copula system consisting of a substantive verb and a proper copula might
have existed in the Brittonic languages, but the verb corresponding to ModIr. tá
and OIr. attá was lost at an early time.
In Brittonic the present and imperfect indicative of the copula is based on Proto-
Celtic *es- from PIE *h1es-, while the forms of all other tenses originate from
PCl. *bu(ː) from PIE *bʰu̯ eh2-.
The substantive verb got lost in the middle period, but the oldest attested texts
evidence its existence. It was based on PCl. *tā-i̯ e/o- from *steh2-. The 9th-
century OW -i-tau (78) and OSWBr. to (80) are both analysed as 3sg indicative
present forms. Since Old British is not as well attested as Old Irish, these isolated
examples do not allow to draw any conclusions about the functions and
frequency of this verb, which is much rarer than the copula is.
In MW, i-taw developed into the subordinating conjunction y taw ‘that’ (79).
OSWBr. to might be contained in the adverb MBr., Br. mar-te-ze ‘perhaps’, if
the latter were the contraction of an expression like *‘if this stands’. The MCo.
interjection otte, at-ta ‘see, behold!’ (81) is analysed as the unclear element ot-,
at- followed by -te or –ta, the 3sg present conjunct of the substantive verb
(Schumacher 2004, 624-6, Fleuriot 1964, 312, Evans/Fleuriot 1985, 559, 561,
Schrijver 2011, 69, 70).
(78) Old Welsh: Juv. fo 11r18, 9th cent.
issit padiu-i-tau gulat gl. celsi thronus est cui regia caeli
COP-I.PRS.3SG to whom-PT-COP-T.PRS.3SG sovereignty
“there is (one), to whom the sovereignty is”20
(79) Middle Welsh: YBH 4.1541, c. 1350 (Schumacher 2011, 216)
ac a dywedwn y taw ti oed Bown
and PT say.IPF.SUBJ.1SG that PRON.2SG be.IPF.3SG Bown
“and I would say that you are Bown”
(80) Old Southwest British: Angers fo 58a, 9th cent. (cited from Fleuriot 1964, 312)
teir trigontoc loir to fiunt
three.FEM of thirty days moon.PL COP-T.PRS.3SG
“it is (of) three moons of thirty days, that they are made”
20
The translation of issit is controversial. Schumacher (2004, 626) compares MW yssit ‘there is’
and translates “Es gibt (einen), dem die Herrschaft ist/zusteht”. Schrijver (2011, 69) analyses iss-id
as ‘there is to him’ = ‘he has’ translating ‘(he) has it, (he) to whom is the kingdom’.
SUPPLETION AND THE COPULA 29
(81) Middle Cornish: OM 117; 1611
Adam, otte an pushes, ydhyn an nef, ha ’n bestes
Adam behold ART fish.PL bird.PL ART heaven and ART animal.PL
“Adam, behold the fish, the birds of the sky and the animals”
8.2 New Forms of the Copula
In all Brittonic languages, the 3sg present indicative has a number of allomorphs,
the use of which is determined by the sentence type (e.g. affirmative, existential,
relative), by properties of the subject (definite or indefinite) or by their position
(clause-initial or not).
One of these allomorphs is MW y mae, MBr. ema and MCo. ymma, which
occurs in the 3rd person present indicative in positive sentences.
MW 3sg (y) mae 3pl (y) maent
MBr ema, e ma emaint, emahint
MCo ma, ym(m)a (y) mons, y myens
Table 7: The new 3rd person present indicative allomorph
The etymology of this word is controversial. As MW (y) mae introduces
affirmative main and relative clauses, Morris Jones (1913, 349) assumed that its
original meaning was ‘there is’ and reconstructed the preform *esmi est
consisting of the locative 3sg masc. of the demonstrative pronoun *e- and the
3sg present indicative of *h1es-.
Hamp (1985, 201) suggested that mae ‘there is’ could have developed from a
locative phrase mae < *maʒe(h)es < *mages-est ‘there is’ containing *mages,
the semantically bleached locative sg. of PCl. *magos ‘plain, field’, cf.
Gaul. -magus, OIr. mag ‘plain’, MW, OSWBr., Corn. -ma ‘place’ and the 3sg
present indicative *est(i). The interrogative expression contained the same
phrase preceded by an interrogative, which seems to have been deleted like in
Welsh pa beth ‘what’ (lit. ‘what thing’) → beth ‘what’.
This reconstruction is not generally accepted. Even though he refers to Hamp’s
work, Schrijver (2011, 69) considers the origin of y mae as problematic.
According to Schumacher (2004, 313) the etymology is unclear. While the 3pl
forms MW y maient and MBr. emaint clearly contain the 3pl of the copula, it
cannot be claimed with certainty that the singular forms MW y mae, MB ema
and MCo. yma contain the corresponding 3sg.
8.2.1 Middle Welsh (y) mae
Middle Welsh (y) mae (COP-M) is used in affirmative and interrogative sentences
in the present tense introducing a main clause or immediately following a(c)
30 BRITTA IRSLINGER
‘and’ introducing a main clause. Complements are locative adverbs (82) or NPs
(83).
When used in questions, (y) mae has the meaning ‘where is...?’ (84). Finally, (y)
mae occurs also in prepositional relative clauses that are introduced by yn ‘in’
and thus are also locative (85).
However, (y) mae cannot be used in negative sentences (86) or if it is preceded
by the subject (90). In these two cases, it is replaced by the copula from *es-,
i.e. yw, ydiw (COP-Y) (Schumacher 2011, 207-216, Evans 1964, 139, 143-144).
Position: (82) PKM 30.14, c. 1350 (Evans 1964, 143)
locative Mae ymma Matholwch brenhin Iwerdon
adverb COP-M.PRS.3SG here Matholwch king Ireland
“Matholwch, the king of Ireland is here.”
Position: (83) BD 151.21f., ms. 14th cent. (Schumacher 2011, 208)
locative NP A y mae llyn arall ... yn emylyeu Kymry
and PT COP.PRS.3SG lake other in border.PL Wales
“and there is another lake ... at the borders of Wales”
Question: (84) PKM 20.23-24, c. 1350 (Evans 1964, 143)
'where is...?' A wraged ... mae y mab?
PT woman.PL COP-M.PRS.3SG ART boy
“Women, where is the boy?”
prepositional (85) WM 119.31f., c. 1350 (Schumacher 2011, 208)
relative clause y lys Arthur yn y mae goreu y gwyr
with yn 'in' in court Arthur in PT COP-M.PRS.3SG best ART man.PL
“at Artus’ court, where the men are best”
Negative (86) Peniarth 14.80.15, 13th cent. (Schumacher 2011, 210)
Nyt ydiw ema.
NEG COP-Y.PRS.3SG here
“He is not here.”
(Y) mae is not restricted to locative uses. It may also denote existence (88),
possession (87) and attribution when used with a noun or a predicative adjective
preceded by the predicative yn (89, 91).21 Finally, the construction y mae + yn +
verbal noun forms the progressive (93, 94).
Again, y mae can neither be used in negative sentences and in tenses other than
the present nor be preceded by the subject or the predicate. In those cases, a form
of the copula based on *es- has to be used (90, 92).
21
This construction becomes frequent in Middle Welsh, cf. Schumacher (2011, 213).
SUPPLETION AND THE COPULA 31
existence (87) PKM 91.16, c. 1350
Ac o achaws hynny y mae digassawc
and because of this PT COP-M.PRS.3SG hostiliy
yr adar y’r tylluan
ART bird.PL to-ART owl.SG
“and therefore there is hostility between the birds and the owl”
possession (88) CO 29.775., ms. c. 1350 (Schumacher 2011, 215)
y mae kerd genhyf i
PT COP-M.PRS.3SG craft with.1SG. PRON.1SG
“I have a craft”
attribution (89) BD 60.29f., ms. 14th cent. (Schumacher 2011, 213)
y mae (+ yn) + ac ar hynny hyt hediw y maent yn ormes
nom. predicate and thereupon until today PT COP-Y.PRS.3PL PRED plague
‘and thereupon they are a plague until today’
attribution (90) PKM 1.1-2 (Evans 1964, 139)
subject first: Pwyll Pendeuic Dyuet, a oed yn arglwyd
yw/ydy (+ yn) + Pwyll prince Dyfed PT COP-Y.IPF.3SG PRED lord
nom. predicate ar seith cantref Dyuet
over seven cantref Dyfed
‘Pwyll, prince of Dyfed, was lord over the seven cantrefs of Dyfed’.
attribution (91) PKM 87.20-21, c. 1350 (Evans 1964, 149)
y mae + yn + a y maent yn barawd
adj. predicate and PT COP-Y.PRS.3PL PRED ready
“and they are ready”
Attribution (92) B v, 215 (Evans 1964, 139) / LlA 109, 1346 (eGPC s.v. parod)
adj. predicate + Parawt wyf i
yw ready COP-Y.PRS.1SG PRON.1SG
“I am ready”
Progressive (93) PKM 82.16, c. 1350
y mae + yn + Ac y maent yn kyrchu y tir
verbal noun and PT COP-M.PRS.3PL PROG make for.VN to Land
“and they are heading for the land”
progressive (94) PKM 29.4 (Schumacher 2011, 212)
imperfect ac yn eisted yd oedynt
and PROG sit.VN PT COP-Y.IPF.3PL
“and they were sitting”
8.2.2 Modern Welsh mae
Modern Welsh mae occurs in the same contexts as Middle Welsh y mae, i.e. it
denotes attribution with a predicative noun or adjective preceded by yn (96, 98),
possession with a possessive phrase (100), existence (101) and position (103)
32 BRITTA IRSLINGER
and it is part of the progressive construction (104). Mae can thus express all
relations with the exception of identity, where only yw/ydy22 can be used (95).
The same is true for negative sentences or cases in which the subject or the
predicate precede the copula (97, 99, 102) (Borsley, Tallerman & Willis 2007,
43, 130f., 258f.).
Identity (95) Borsley, Tallerman & Willis (2007, 130)
NO mae Prifddinas Cymru yw Caerdydd.
capital Wales COP-Y.PRS.3SG Cardiff
“Cardiff is the capital of Wales.”
Attribution (96) Borsley, Tallerman & Willis (2007, 43)
with NP Mae Gwyn yn feddyg.
COP-M.PRS.3SG Gwyn PRED doctor
“Gwyn is a doctor.”
Attribution (97) Borsley, Tallerman & Willis (2007, 256)
predicate first: Athro ydy Gwyn.
copula yw/ydy teacher COP-Y.PRS.3SG Gwyn
“Gwyn is a teacher.”
Attribution (98) Borsley, Tallerman & Willis (2007, 43)
with adjective Mae Gwyn yn ddiog.
COP-M.PRS.3SG Gwyn PRED lazy
“Gwyn is lazy.”
attribution (99) Borsley, Tallerman & Willis (2007, 256)
predicate first: Diog ydy Gwyn.
copula yw/ydy lazy COP-Y.PRS.3SG Gwyn
“Gwyn is lazy.”
possession (100) Borsley, Tallerman & Willis (2007, 63)
Mae car newydd gan Megan
COP-M.PRS.3SG car new with Megan
“Megan has a new car.”
existence (101) Borsley, Tallerman & Willis (2007, 257)
Mae ceffyl yn yr ardd.
COP-M.PRS.3SG horse in ART garden
“There is a horse in the garden.”
existence (102) Borsley, Tallerman & Willis (2007, 257)
negative Nid oes ceffyl yn yr ardd.
NEG COP-Y.PRS.3SG horse in ART garden
“There isn't a horse in the garden.”
22
Ydy is the northern variant of the copula and yw the southern one.
SUPPLETION AND THE COPULA 33
Position (103) Borsley, Tallerman & Willis (2007, 43)
Mae Gwyn dan y bwrdd.
COP-M.PRS.3SG Gwyn under ART table
“Gwyn is under the table.”
progressive (104) Borsley, Tallerman & Willis (2007, 248)
mae + yn + Mae Gwyn yn cysgu.
infinitive COP-M.PRS.3SG Gwyn PROG sleep.INF
“Gwyn is sleeping.”
In light of what was just explained, which conclusions are to be drawn about the
function of mae? If the etymology is correct and mae was originally a positional
verb, it means that over time it has taken over all functions of the copula except
the expression of identity. Nevertheless, nominal and adjectival attributes only
occur with predicative yn, thus constrasting with the construction used with
yw/ydy.
Rouveret (1996, 133) attempts to interpret mae-constructions as stage level
predicates, even when they contain individual level predicates. According to
him, the adjective glas in ex. (105) is an individual level predicate turned into
the stage level predicate yn las23 by the marker yn.
attribution (105) Rouveret (1996, 133)
permanent attribute Mae ’r mor yn las.
COP-M.PRS.3SG ART sea PRED blue
“The sea is blue.”
Hendrick (1996, 88) reasonably rejects this analysis arguing that the distribution
of mae is not influenced by the semantic character of the predicate.
As the preceding survey has shown, only syntactic features are relevant to the
use of mae. Mae is used in positive sentences in the present indicative with two
constraints: 1. it may not occur with a definite complement or a complement
expressing identity and 2. it may not be preceded by its complement. These
constraints contributed to the bleaching of the locative semantics which (y) mae
may once have possessed if Hamp’s etymology is correct.
8.2.3 Middle Breton ema
Another development is found in Middle Breton, where the marked copula has
the form ema (also e ma, ez ma)24 and indicates the position of a definite subject
23
The adjective glas is mutated after yn.
24
The verbal particle e, which corresponds to MW y, fuses increasingly with ma and is not analysed
as a verbal particle in synchronic descriptions of Modern Breton. Like the other verbal particles it is
frequently dropped in spoken Breton.
34 BRITTA IRSLINGER
in clauses with a locational complement (106). The same construction can be
found in more metaphoric expressions (107), denoting a situation rather than a
concrete location. Like in Welsh, it cannot be used after a nominal or pronominal
subject. Sentences starting with this type of subject, i.e. original cleft sentences,
use the relative form of the unmarked copula (108) (Schrijver 2011a, 418,
Hemon 1975, 205, GIB s.v. bezañ, Meurgorf s.v. bezañ).
position (106) Do., p. 44; 1622 (cited from Meurgorf s.v. bezañ)
subject after verb ez ma eno corff hon Saluer Biniguet
PT COP-M.PRS.3SG in.3SG.M body POSS.1PL saviour blessed
“in it is the body of our blessed saviour”
situation (107) LDJM 1; 1659 (Meurgorf s.v. bezañ)
(temporary) ema e pirill
PT.COP-M.PRS.3SG in danger
“he is in danger”
position (108) corff hon Saluer Biniguet a zo eno
subject before verb body POSS.1PL saviour blessed PT COP.REL in.3SG.M
“the body of our blessed saviour is in it”
8.2.4 Modern Breton emañ
From the middle of the 17th century onwards, finite forms arose to complete the
paradigm, especially in the dialect of Léon. These can also be used in negative
sentences (115), although some dialects (Vannetais, Pélem, Eastern
Cornouaille) prefer the unmarked copula eo. (Favereau 1997, 233-236, Hemon
1975, 205-6, 268f.). Some dialects also display an imperfect paradigm build on
the stem ed- in Léon and em-ed- in Cornouaille and in parts of Léon (Favereau
1997, 211, Hemon 1975, 206, Hemon 1954, 224-227).
1S emaon 1P emaomp
2S emanout 2P emaoc'h
3S emañ 3P emaint
Impersonal emeur
Table 8: Inflection of emañ, present indicative
Emañ is used to indicate position in locative (109) and in temporal contexts
(110) with the according NPs or adverbs and, additionally, it denotes more
abstract situations (111-114).
position (109) GReg.; 1732 (Meurgorf s.v. bezañ)
locative e ma ê Kaer
PT COP-M.PRS.3SG in town
“he is in town”
SUPPLETION AND THE COPULA 35
Temporal (110) Favereau (1997, 223)
d’al lun ema foar Rostren’n
to-ART Monday PT.COP-M.PRS.3SG fair Rostrenen
“on Moday is the fair of Rostrenen”
Situation (111) GReg, pg abbois, 1732 (cited from Meurgorf s.v. bezañ)
ema ar maro gantâ
PT.COP-M.PRS.3SG ART dead with.3SG.M
“dead is with him”
(112) SVBV p. 10;1929 (cited from Meurgorff s.v. bezañ)
Mar d-eo evel-se eo emañ
if COP-E.PRS.3SG like this COP-E.PRS.3SG PT.COP-M.PRS.3SG
an traou
ART thing.PL
“if things are like this” (lit.: “if it is like this that things are”)
(113) RP p. 42; 1718 (cited from Hemon 1975, 268-9)
d’ar poent ma ema o rei
to-ART time when PT.COP-M.PRS.3SG PROG give.INF
deomp peadra da veva
to.1PL enough to live.INF
“in a time when he is giving us enough to live”
with participle (114) FG p. 67; 18th cent. (cited from Hemon 1975, 247)
e ma aet Margod kuit
PT COP-M.PRS.3SG gone Margot away
“Margot has gone away.”
negative, with (115) Faverau (1997, 235)
infinitive n’ ema ket da werzhañ
NEG PT.COP-M.PRS.3SG NEG to sell.INF
“It is not for sale”
With a nominal or adjectival attribute, emañ indicates a temporary state (116,
118), whereas the unmarked copula eo presents the statement as a matter of fact
(117, 119). Finally, emañ may occur in a progressive construction with the
progressive particle o + infinitive (113, 120) (Favereau 1997, 233-6).
nominal attribute (116) Favereau (1997, 234)
momentary setu ‘mañ brav an amzer
situation now COP-M.PRS.3SG nice ART weather
“now the weather gets nice.”
intrinsic quality (117) brav eo an amzer
nice COP-E.PRS.3SG ART weather
“the weather is nice”/”this is nice weather”
36 BRITTA IRSLINGER
adejctival attribute (118) Favereau (1997, 234)
temporary state ’ma klañv
COP-M.PRS.3SG ill
“he/she is ill (at the moment)”
matter of fact (119) klañv (ez) eo
ill PT COP-E.PRS.3SG
“he/she is ill”
progressive (120) Favereau (1997, 234)
with o + infinitive O labourat emaoc’h?
PROG work.INF COP-E.PRS.2PL
“Are you working?”
In contrast to Welsh, Breton does have a special copula for locational
predication, which possesses a complete paradigm and can be used in negative
sentences.25
9. Evaluation of the Evidence
Vennemann and Trudgill built their contact based hypothesis on the fact that the
two-copula systems with the same aspectual distinctions are attested in
neighbouring languages:
“So the existence in the same geographical area of Europe of a two-copula system, with a
distinction precisely and perhaps unusually between a habitual and a non-habitual copula, in
members of three different language families, does seem to signal some kind of contact-based
explanation.” (Trudgill 2011, 5)
However, the present survey showed serious objections on the basis of a detailed
examination of the systems in question. The question of how the aspectual
oppositions should be evaluated remains unanswered. Furthermore, the
substratum hypothesis is less plausible if the underlying developments are cross-
linguistically widespread.
9.1 Aspectual Oppositions
The copula systems of the languages in question show two distinct oppositions
correlated with different roots. While Middle Welsh and Old English share the
habitual : non-habitual distinction expressed by roots based on PIE *h1es- and
*bʰu̯ eh2-, the Romance and Irish copulas originate from PIE *h1es- and *steh2-
and, at least to a certain degree, express contrast between permanent and
temporary attributes or individual level vs. stage level predicates respectively.
25
Stassen’s (2005, 484, map 119) classification of Welsh and Breton with regard to nominal and
locational predication is thus erroneous. While Breton possesses split or different encoding for
nominal and location predication, in Welsh the encoding is shared or identical.
SUPPLETION AND THE COPULA 37
The same semantic distinctions can also be found in Basque. To prove that the
majority of Western European languages possess two-copula systems with
identical differentiations, both Vennemann and Trudgill (2011, 4) equate the
habitual : non-habitual distinction with the permanent : temporary one.
Both distinctions have been considered as “grammatical aspect” or as “lexical
aspect”, i.e. Aktionsart ‘actionality’, neither of which have received universally
accepted definitions (Binnick 2001, 561). Comrie (1976, 25f.) considers the
Welsh habitual as a subtype of imperfective aspect, while he discusses the Ibero-
Romance and Irish copulas in the chapter “contingent state” (p. 103).
Carlson (2012, 830ff.) argues against the equivalence of individual level
expressions and habitual and generic sentences. He writes a dedicated section
for the Spanish copulas, i.e. “statives” (p. 839-41), rather than dealing with them
in the section about “habitual aspect” (p. 831-2). The reason being that “statives
generally do not appear to participate in habituality paradigms”, although there
is a certain semantic overlap with regard to the differentiation between
temporary and more long-term states.26
Despite the fact that these theoretical issues need to be researched further, the
separation of the two concepts is adequate in view of the complex Insular Celtic
systems. Both Irish and Breton actually possess not only two, but three different
copulas.
[nominal] [locational] [habitual]
Irish is tá bíonn
Breton eo (a zo, eus) emañ bez
Welsh yw/ydy (mae, oes) – bydd
Table 9: Aspectual distinctions of copulas, 3sg present indicative
In Modern Irish, the habitual present and past are used in the same contexts as
the substantive verb to denote habitual or iterative actions (ex. 121). The copula,
which in any case denotes more permanent states, does not have a habitual
counterpart.
(121) NIG 92
Bím ar scoil gach lá.
COP-T.HAB.PRS.1SG at school every day
“I am at school every day.”
26
Maienborn (2005, 171-175) argues that the ser/estar alternation is not aspectual but rather a
pragmatic phenomenon, as estar occurs in arbitrary topic situations and ser in specific ones.
38 BRITTA IRSLINGER
The habitual is already present in Old Irish and here it can also be found with
the copula, cf. ex. (122) featuring a nominal and an adjectival predicate.27 The
habitual form of the copula is attested without doubt only for the 3rd persons, but
Schumacher (2009, 256, fn. 38) assumes that the paradigm was complete.
(122) legal maxim, CIH V 1608.14, translation by Schumacher (2009, 256)
Níbi briugu nádbi cétach.
NEG.COP-I.HAB.PRS.3SG hospitaller.NS NEG.REL.COP-I.HAB.PRS.3SG possessing a hundred
“He is generally not a briugu (hospitaller) who does not have hundred-fold possessions.”
Therefore, the substantive verb attá has the habitual present biid ‘is wont to be,
is continually’. Analogously, the copula is has bid (with short i) as its habitual
counterpart, although is can express habituality as well (GOI 331, 487).
aspect
non-habitual habitual/repeated
preterite imperfect PCl. present stem *bu̯ -ii̯ e/o- ‘to become, grow’
tense present — > future/habitual: ‘to become / to be usually’
Table 10: Development of the Insular Celtic habitual present
Repeated or customary actions are expressed in the past tense by the imperfect,
an inflectional category inherited from PIE. This category is open not only to
the copula, but to all verbs (GOI 331). In the past tense there was thus an
inflectionally expressed distinction between habitual and non-habitual, which
had no parallel in present tense.28 The present stem *bu̯ -ii̯ e/o-29 of PIE *bʰu̯ eh2-
‘to become, grow, come into existence’ stepped in as the future/habitual ‘to
become / to be usually’.
9.2 STAND as a Copula in IE Languages
Cross-linguistically, the majority of languages encodes nominal predication
differently than locational predication. Almost 70 % of the languages of
Stassen’s (2005, 482) sample belong to this type. Although the encoding
strategy for locational predication consists not necessarily in a position verb,
27
In ex. (120), the distinction between the habitual and non-habitual seems somewhat contrived.
Briugu ‘hospitaller’ is one of the ranks of medieval Irish personal law. This rank is conceived as a
permanent state. Kelly (1988, 36) does thus not use ‘usually’ in the translation of this passage: “He
is not a briugu who is not a possessor of hundredfold wealth”. Schumacher’s examples of the
habitual use of the copula are all from legal texts and might be genre-specific.
28
According to Bybee/Perkins/Pagliuca (1994, 154) special habitual markers are cross-linguistically
more frequent in past tense than in present tense, since explicit reference to habitualness is necessary
in the past, but not in the present.
29
See Schumacher (2004, 247) for the developments underlying the individual forms.
SUPPLETION AND THE COPULA 39
position verbs or, more precisely, the bodily posture verb STAND together with
SIT and LIE as copulas and auxiliaries are also cross-linguistically frequent,
especially as continuative / durative / progressive markers (Kuteva 1999, Ströbel
2010).
For the Latin position verbs Stengaard (1991, 14-20) established three semes or
semantic properties, each of which may be more or less dominant.
(123) position/bodily posture [pos] + secondary features, e.g. [resultative]
duration [dur] [cumulative]
location/spatial position [loc]
Postural verbs originally denote the bodily posture of humans in space. During
the grammaticalisation process, their scope is extended to express the spatial
position of inanimates. The feature of “unboundedness” is inherently present in
their meaning, but while it is non-focal in stage I, it becomes more focal during
the grammaticalisation process and may ultimately develop into a continuative
/ durative / progressive aspectual marker, eventually acquiring additional
features (Kuteva 1999, 206-8). Alternatively, the locational element may
become focal and be associated with the actual discourse situation (Maienborn
2005, 174). Related concepts are the expression of existence (‘be present, exist’)
or temporary properties.
For copulas, Remberger/Gonzáles-Vilbazo (2007, 207) established the
following grammaticalisation path:
postural verb → locative verb → existential verb → copula with a temporary sense
The grammaticalisation of STAND into a copula or auxiliary is thus not
restricted to Western European languages. Remberger / Gonzáles-Vilbazo
(2007, 216-7) add Neapolitan and Kuteva (1999, 206-8) discusses, among
others, examples from Old Dutch and Bulgarian. Matching developments are
also found in Vedic, an Indo-European language which is clearly situated
outside the area of alleged Celtic or Vasconic influence. Lühr (2007, 192-5)
adduces examples from the Rig-Veda where the verbs sat ‘to sit’ (from PIE
*sed-) or sthā ‘to stand’ (from PIE *steh2-) are constructed with a present
participle (or a perfect participle with present meaning) in a way that is
equivalent to an English gerund construction. As the original postural semantics
of both verbs got bleached while the durative property became focal, Lühr
suggests to translate both of them as ‘to be busy, engaged, occupied’, cf. ex.
(124) featuring a form of sthā:
40 BRITTA IRSLINGER
(124) RV 1,35,10 (Lühr 2007, 194)
apasédhan rakṣáso yātudhā ́ nān ásthād deváḥ pratidoṣáṃ
keep off.PTC.PRS.NS.M demon.AP sorcerer.AP COP.IND.AOR.3S god.NS evening.ADV
gṛṇānáḥ
praise.PTC.PRS.MID.NS.M
“In the evening, the god is busy to keep off the demons and sorcerers, the divine one.”
(Geldner: “Die Unholde und Zauberer abwehrend steht der Gott allabendlich da.”)
In Hindi, the verb sthā became part of the paradigm of the auxiliary honā ‘to
be’. The present tense of honā consists of forms belonging to the root as (from
PIE *h1es-), while the imperfect is supplied by suffixed forms of the verbal
adjective of the root sthā. Together with the participle I (based on the old present
participle), the auxiliary thā/thī forms the periphrastic imperfect, which
expresses duration (Oberlies 1998, 29).
present tense imperfect
Hindi Sanskrit masc. fem.
1S hūṁ < *asāmi thā < *sthitakaḥ thī < *sthitakā
2S hai < *asasi thā thī
3S hai < *asati thā thī
1P haiṁ the < *sthitakāḥ thīṁ < *sthitakāni
2P ho < *asatha the thīṁ
3P haiṁ the thīṁ
Table 11: Inflection of Hindi honā ‘to be’ in present and imperfect indicative
10. Conclusions
This diachronic and typological survey of the copula systems of Insular Celtic,
Romance and Basque yielded the following results:
Spanish, Portuguese, Catalan, the Western Basque dialects and Irish employ a
special copula for locational predicates, which, in the case of the IE languages,
is based on PIE *steh2- ‘to stand’. Within the individual languages, this copula
has reached different degrees of grammaticalisation at different times.
The Old Irish copula attá, which does not express bodily posture, indicates
location and existence. From the 11th century onwards, it appears in a locative
construction expressing temporary properties with nominal predicates.
The Spanish, Portuguese and Catalan two-copula systems, in which ser and
estar are used to contrast permanent and temporary attributes, developed from
the 13th centuries onwards. There estar spread from locational to adjectival and
finally to nominal predicates, whereas in Portuguese only the latter of these
stages has developed.
SUPPLETION AND THE COPULA 41
Contrary to the findings of earlier research, neither Vulgar Latin nor Old French
have clear evidence for STAND in the function of a copula. In the case of Old
French, it remains even doubtful whether Latin stāre underlies the new
imperfect that replaced the inherited forms from the 13th century onwards.
While the Western Basque dialects have a two-copula system of the Portuguese
type, the Eastern Basque dialects have a one-copula system like French. Due to
the lack of evidence, the developments that lead to this systems cannot be traced
or dated.
Although the grammaticalisation of STAND in neighbouring languages can still
be interpreted as an areal feature, its occurrence from the late Middle Ages
onwards makes a pre-Indo-European substratum as its cause highly improbable.
A possible Celtic influence on Gallo-Latin would be also difficult to prove
(leaving aside the fact, that the copula system of Gaulish is unknown).
Alternatively, it could simply be a coincidence that Irish and several Romance
languages have reached a similar degree of grammaticalisation roughly at the
same time, while equivalent grammaticalisation processes were going on in the
neighbouring languages. In addition, it has so far gone unnoticed that Breton
developed a new locative copula after the Brittonic languages had lost the verb
based on *steh2- shortly after the beginning of the written transmission.
Outside the area of Western Europe, Vedic sthā has reached a high degree of
grammaticalisation at an early time. Likewise, forms of *steh2- were integrated
into the paradigm of *h1es- in Tocharian and Hindi.
The Middle Welsh and Old English copula systems, on the other hand, do not
seem to related to the Irish and Ibero-Romance developments. Although the
details of the process remain unclear, there is a high probability for Celtic
influence on the Old English copula system. In addition, another root and
function are involved, i.e. PIE *bʰu̯ eh2- marking habitual aspect.
In conclusion, the hypothesis that the copula systems of most Western European
languages results from the same substratum influence cannot been confirmed.
42 BRITTA IRSLINGER
Gloss abbreviations
art definite article neg negator
abs absolutive on nominals m masculine
adv adverb mid middle
aor aorist n neuter
ap accusative plural nom nominalizer
as accusative singular ns nominative singular
comp complementizer pl plural
cond conditional ptc participle
cop copula poss possessive
dem demonstrative pred predicative
dep dependent prep preposition
det determiner (article) prev preverb
emph emphatic prog progressive
f feminine pron pronoun
fut future prs present
gs genitive singular prs.hab habitual present
imp imperative prt preterite
ind indicative pt particle
inf infinitive rel relative
ips impersonal ending sg singular
ipf imperfect subj subjunctive
loc locative vn verbal noun
Language abbreviations
O Old e.g. OIr. Old Irish Gaul. Gaulish
M Middle e.g. MW Middle Welsh Ir. Irish
Mod Modern e.g. ModBr. Modern Breton SWBr. Southwest British
Br. Breton W Welsh
Co. Cornish PCl. Proto-Celtic
E English PIE Proto-Indo-European
Fr. French
References
Primary Sources
ACC: Stokes, Whitley (ed. and tr.) (1899): The Bodleian Amra Choluimb chille. Revue Celtique 20:
31–55, 132–183, 248–289, 400–437. Corrigenda in Revue Celtique 21 (1900): 133–136.
Barl. u. Jos.: Appel, Carl (ed.) (1907): Gui v. Cambrai: Balaham und Josaphas. Nach den Hss. von
Paris und Monte Cassino hrsg. v. Carl Appel. Halle.
BD: Lewis, Henry (ed.) (1942): Brut Dingestow. Llandysul.
Beowulf: Fulk, Robert D., Bjork, Robert E. & Niles, John D. (eds.) (2008): Klaeber’s Beowulf and
the fight at Finnsburg. Ed., with introduction, commentary, appendices, glossary, and
bibliography. 4. ed., based on the 3. ed. with 1. and 2. suppl. Toronto/Buffalo/London.
Cantar de mio Cid: Menéndez Pidal, Ramón (ed.) (1955): Poema de mio Cid: vers 1140. Texte
critique établi par Ramón Menéndez Pidal. Trad. française et préf. de Eugène Kohler. Paris.
SUPPLETION AND THE COPULA 43
CO: Bromwich, Rachel. (ed.) (1992): Culhwch and Olwen. An edition and study of the oldest
Arthurian tale. Cardiff.
Do.: Doctrin Christen. Morlaix 1622.
FB: Henderson, George (ed. and tr.) (1899): Fled Bricrend : = The Feast of Bricriu. An early Gaelic
saga transcribed from older mss. into the Book of the Dun Cow by Moelmuiri Mac Mic Cuinn na
M-Bocht. Ed., with transl., introd. and notes. London.
FG: Kerenveier, Paskal (18th cent.): Ar Farvel Goapaer. Comedy published in Sterenn 6 (Brest
1941).
Genesis: Crawford, S. J. (ed.) (1922): The Old English version of the Heptateuch. Aelfric’s Treatise
and the Old and New Testament and his preface to the Genesis. London.
GReg: De Rostrenen, Grégoire (1732): Dictionnaire françois-celtique ou françois-breton. Rennes.
Juv.: McKee, Helen (ed.) (2000): The Cambridge Juvencus manuscript. Glossed in Latin, Old
Welsh, and Old Irish. Text and commentary. Aberystwyth.
LDJM: Le Menn, Gwennole (ed.) (1996): Les dictionnaires français-breton et breton français du
R.P. Julien Maunoir. 1659. En 2 tomes, avec index géneral. Réédité et annoté par Gwennole Le
Menn. [Saint-Brieuc].
LlA: Jones, John Morris / Rhŷs, John (eds.) (1894): The Elucidarium and OtherTtracts in Welsh
from Llyvyr agkyr Llandewivrevi A.D. 1346 (Jesus college ms. 119). Oxford.
OM: Harris, Phyllis Pier (ed.): Origo mundi, First Play of the Cornish Mystery Cycle, the Ordinalia.
University of Washington dissertation 1964.
Peniarth 14: Rhyddiaith y 13eg Ganrif: Fersiwn 2.0. A Searchable Version of All the Welsh Prose
in 13th-Century Manuscripts. Transcribed by G. R. Isaac & Simon Rodway (2002) and
reformatted by Silva Nurmio, Kit Kapphahn & Patrick Sims-Williams (2010); emended by Simon
Rodway & Patrick Sims-Williams (2013) © Department of Welsh, Aberystwyth University 2010,
2013. http://cadair.aber.ac.uk/ dspace/handle/2160/11163
PH: Atkinson, Robert (ed., tr.) (1887): The Passions and Homilies from Leabhar Breac. Text,
translation, and glossary by Robert Atkinson. Dublin.
PKM: Williams, Ifor (ed.) (1930): Pedeir Keinc y Mabinogi. Caerdydd.
Queste: Pauphilet, Albert (ed.) (1923): La queste del saint Graal. Roman du XIIIe siècle. Paris.
R: Gwenogvryn Evans, John (ed.) (1911): The Poetry in the Red Book of Hergest. Llanbedrog.
RP: Bris, Ch. (c. 1718): Reflexionou profitabl var ar Finvezou diveza.
SVBV: Ar Gall, Yann (Ab-Sulio) (1929): Seiz vloaz e bro ar Vorianed. Brest.
TBC I: O'Rahilly, Cecile (1976): Táin Bó Cúailnge. Recension I. Dublin.
TBC II: O'Rahilly, Cecile (1967). Táin Bó Cúalnge. From the Book of Leinster. Dublin.
TBDD: Stokes, Whitley (ed.) (1901-02): Togail bruidne Da Derga. The destruction of Da Derga's
hostel. Revue Celtique 22: 9-61, 165-215, 282-329, 390-437; 23, 88.
Thes.: Stokes, Whitley / Strachan, John (eds.) (1975): Thesaurus Palaeohibernicus. A collection of
Old Irish glosses and scholia prose and verse. 2 vols. Repr. Dublin.
Ven. Fortunatus: Leo, Fridericus (ed.) (1881): Venanti Honori Clementiani Fortunati Presbyteri
Italici Opera Poetica. Beolini.
Fels, Wolfgang (tr.) (2006): Venantius Fortunatus: Gelegentlich Gedichte. Eingeleitet, übersetzt
und kommentiert von Wolfgang Fels. Stuttgart.
Wb.: The Würzburg glosses on the Pauline epistles, cf. Thes. i, 499ff.
WM: Gwenogvryn Evans, John (ed.) (1907): The White Book Mabinogion. Welsh tales and
romances reproduced from the Peniarth manuscripts. Pwllheli.
YBH: Watkins, Morgan (ed.) (1958): Ystorya Bown o Hamtwn. Cyfieithiad canol y drydedd ganrif
ar ddeg o La geste de Boun de Hamtone. Caerdydd.
44 BRITTA IRSLINGER
Secondary Sources
Adams, Douglas Q. (2013): A Dictionary of Tocharian B. 2 vols. Second edition, revised and greatly
enlarged. Amsterdam ; New York, NY.
Baldi, Philip / Page, Richard (2006): Review of "Europa Vasconica – Europa Semitica". Lingua
116.12: 2183-2220.
Binnick, Robert I. (2001): 42. Temporality and aspectuality. In: Haspelmath, Martin / König,
Ekkehard / Oesterreicher, Wulf / Raible, Wolfgang (eds.): Language Typology and Language
Universials. An International Handbook I.1. Berlin / New York: 557-567.
Bonnard, Henry / Regnier, Claude (1989): Petite grammaire de l'ancien français. Paris.
Borsley, Robert D. / Roberts, Ian (eds.): The Syntax of the Celtic Languages. A Comparative
Perspective. Cambridge.
Borsley, Robert D. / Tallerman, Maggie / Willis, David (2007): The Syntax of Welsh. Cambridge.
Bouzet, Jean (1953): Orígines del empleo de ‘estar’. In: Estudios dedicados a Menéndez Pidal Vol.
IV. Madrid. 37-58.
Bybee, Joan / Perkins, Revere / Pagliuca, William (1994): The Evolution of Grammar. Tense, Aspect,
and Modality in the Languages of the World. Chicago and London.
Carlson, Greg (2012): Habitual And Generic Aspect. In: Binnick, Robert I. (ed.): The Oxford
Handbook of Tense and Aspect. Oxford: 828-851.
Cano Aguilar, Rafael (2002): El español a través de los tiempos. 5. ed. Madrid.
Clancy, Steven J. (2010): The Chain of “Being” and “Having” in Slavic. Amsterdam.
Comrie, Bernard (1976): Aspect. An introduction to the study of verbal aspect and related problems.
Cambridge.
De Bruyne, Jacques (1995): A Comprehensive Spanish Grammar. Adapted with aditional material
by Christpher J. Pountain. Oxford.
De Lage, Guy Raynaud (1990): Introduction à l'ancien français. Nouv. éd. par Geneviève Hasenohr.
Paris.
De Rijk, Rudolf P.G. (2008): Standard Basque. A Progressive Grammar. Vol. 1: The Grammar.
Vol. 2: The glosses. Prepared by Armand De Coene. Cambridge, Mass.
Dillon, Myles (1927): Nominal Predicates in Irish I. Zeitschrift für celtische Philologie 16: 313-356.
Dillon, Myles (1928): Nominal Predicates in Irish II. Zeitschrift für celtische Philologie 17: 307-
346.
Dixon, Robert M. W. (2010): Basic Linguistic Theory. Vol. 2: Grammatical topics. Oxford.
eDIL 2013: Dictionary of the Irish Language: Based Mainly on Old and Middle Irish Materials.
Dublin 1913-76. (http://edil.qub.ac.uk/dictionary/search.php).
Ernout, Alfred / Meillet, Antoine (1979): Dictionnaire etymologique de la langue latine. Histoire
des mots. 4. ed. augm. d'add. et de corr. nouv. par Jacques Andre. Paris.
Evans, Claude / Fleuriot, Léon (1985): Dictionnaire du vieux breton / A Dictionary of Old Breton:
Historical and Comparative. Vol. II: A Supplement, Including an English Translation by Claude
Evans of an Abridged Version of the Dictionary, and Additional Material. Toronto.
Evans, Simon D. (1964): A Grammar of Middle Welsh. Repr. 1989. Dublin.
Favereau, Francis (1997): Grammaire du breton contemporain / Yezhadur ar brezhoneg a-vremañ.
Morlaix.
Fleuriot, Léon (1964): Dictionnaire des gloses en vieux breton. Paris.
GG = The Christian Brothers (1999): Graiméar Gaeilge na mBráithre Críostaí. Baile Átha Cliath.
GIB = Hemon, Roparz (1979–1998): Geriadur istorel ar brezhoneg / Dictionnaire historique du
breton. 2. éd. Plomelin.
GOI: Thurneysen, Rudolf (1946): A Grammar of Old Irish. With supplement. Revised and enlarged
Transl. from the German by D. A. Binchy and Osborn Bergin. Dublin.
GPC: R. J. Thomas et al. (eds.). 1950 –: Geiriadur Prifysgol Cymru. A dictionary of the Welsh
language. Caerdydd. http://geiriadur.ac.uk/gpc/gpc.html.
SUPPLETION AND THE COPULA 45
Hamp, Eric P. (1985): XXIII. Welsh mae 'Where is?'. Études Celtiques 22: 210.
Heine, Bernd / Kuteva, Tania (2002): World Lexicon of Grammaticalisation. Cambridge.
Heine, Bernd / Kuteva, Tania (2005): Language Contact and Grammatical Change. Cambridge.
Hemon, Roparz (1954): On Some Forms of the Verb ‘to Be’ in Breton. Celtica 2: 217-228.
Hemon, Roparz (1975): A Historical Morphology and Syntax of Breton. Dublin.
Hendrick, Randall (1996): Some Syntactic Effects of Suppletion in the Celtic Copulas. In: Borsley
/ Roberts (eds.): 75-96.
Hengeveld, Kees (1986): Copular Verbs in a Functional Grammar of Spanish. Linguistics 24: 393-
420.
Hengeveld, Kees (1991): Tipología, sincronía, diacronía. In: Haverkate, Hank / Hengeveld, Kees /
Olbertz, Hella (eds.): Exploracíones semánticas y pragmáticas del Español. Amsterdam: Rodopi:
81-94.
Hengeveld, Kees (1992): Non-verbal Predication. Theory, Typology, Diachrony. Berlin / New York.
Hickey, Leo D. (1968): Los verbos de existencia en castellano e irlandes. Filologia moderna 8: 213-
227.
Hickey, Raymond (ed.) (2010): The Handbook of Language Contact. Oxford.
Hualde, José Ignacio / Lakarra, Joseba A. / Trask, Robert L. (eds.) (1995): Towards a History of the
Basque Language. Amsterdam / Philadephia.
Hualde, José Ignacio / Ortiz De Urbina, Jon (eds.) (2003): A Grammar of Basque. Berlin.
Juge, Matthew L. (2000): On the Rise of Suppletion in Verbal Paradigms. Proceedings of the 25th
Annual Meeting of the Berkeley Linguistics Society: 183-194.
Keller, Wolfgang (1925): Keltisches im englischen Verbum. In: Anglica. Untersuchungen zur
englischen Philologie. Alois Brandl zum siebzigsten Geburtstage überreicht. Bd. I: Sprache und
Kulturgeschichte. Leipzig: 55-66.
Kelly, Fergus (1988): A Guide to Early Irish Law. Dublin.
Kölligan, Daniel (2007): Suppletion und Defektivität im griechischen Verbum. Bremen.
Kuteva, Tania A. (1999): On ‘sit’/‘stand’/‘lie’ auxiliation. Linguistics 37.2: 191-213.
Lapesa, Rafael (2008): Historia de la lengua española. Prólogo de Ramón Menéndez Pidal. Madrid.
LIV: Rix, Helmut et al. (2001): Lexikon der indogermanischen Verben. 2. ed. Wiesbaden.
Löfstedt, Einar (1911): Plautinischer Sprachgebrauch und Verwandtes. Glotta 3.2: 171-191.
Lutz, Angelika (2010): Celtic Influence on Old English and West Germanic. English Language and
Linguistics 13.2: 227-249.
Maienborn, Claudia (2005): A Discourse-based Account of Spanish ser/estar. Linguistics 43.1: 155-
180.
Mateus, Maria Helena Mira / Brito, Ana Maria / Duarte, Inês Silva / Faria, Isabel Hub (1983):
Gramática da língua portuguesa. Elementos para a descrição da estrutura, funcionamento e uso
do português actual. 1. ed. Coimbra.
McCone, Kim (2005): A First Old Irish Grammar and Reader. Including an Introduction to Middle
Irish. Maynooth.
Meurgorf: Dictionnaire historique Meurgorf. http://meurgorf.brezhoneg.bzh/
Meyer-Lübke, Wilhelm (1935): Romanisches etymologisches Wörterbuch. 3., vollst. neubearb.
Aufl. Heidelberg.
Morris Jones, John (1913): A Welsh Grammar. Historical and Comparative. Oxford.
NIG: The Christian Brothers (2002): New Irish Grammar. Repr. Naas, Co. Kildare.
Nolan, Brian (2012): The Structure of Modern Irish. A Functional Account. Sheffield.
Ó Siadhail, Mícheál (1989): Modern Irish. Grammatical Structure and Dialectal Variation. Cam-
bridge.
Oberlies, Thomas (1998): Historische Grammatik des Hindi. Die Genese seines morphologischen
Systems aus dem Mittel- und Altindischen. Reinbek.
46 BRITTA IRSLINGER
Ortiz De Urbina, J. 2010: Periphrastic contructions. In: Hualde / Ortiz de Urbina (eds.) (2010): 285-
300.
Penny, Ralph (2000): Variation and Change in Spanish. Cambridge.
Perini, Mário A. (2002): Modern Portuguese. A Reference Grammar. New Haven / London.
Petré, Peter (2013): On the Distribution and Merger of Is and Bið in Old and Middle English.
Transactions of the Philological Society 111.3: 301-325.
Pinault, Georges-Jean (1990): Introduction au Tokharien. In: LALIES. Actes des sessions de
linguistique et de littérature 7. Paris. 1-224.
Pokorny, Julius (1959): Keltische Urgeschichte und Sprachwissenschaft. Die Sprache 5: 152-164.
Pokorny, Julius (1962): Die Sprachen der vorkeltischen Bewohner Nordwesteuropas. In: Knobloch,
Johann (Hrsg.): II. Fachtagung für indogermanische und allgemeine Sprachwissenschaft.
Innsbruck, 10.-15. Oktober 1961. Innsbruck: 129-138.
Pountain, Christopher (1982): *essere/stare as a Romance Phenomenon. In: Vincent, Nigel / Harris,
Martin (eds.): Studies in the Romance Verb. Essays Offered to Joe Cremona on the Occasion of
his 60th Birthday. London / Canberra: 139-160.
Pountain, Christopher (1985): Copulas, Verbs of Possession and Auxiliaries in Old Spanish: the
Evidence for Structurally Interdependent Changes. Bulletin of Hispanic Studies 62: 337-355.
Pustet, Regina (2003): Copulas. Universals in the Categorization of the Lexicon. Oxford.
Ramos Alfajarín, Joan Rafael (2000): Ésser, estar i haver-hi en català antic. Estudi sintàctic i
contrastiu. València.
Remberger, Eva / González-Vilbazo, Kay-Eduardo (2007): Die Kopula im Romanischen. In: Geist,
Ljudmila / Rothstein, Björn (eds.): Kopulaverben und Kopulasätze. Intersprachliche und
intrasprachliche Aspekte. Tübingen: Niemeyer, 201-226.
Revol, Thierry (2000): Introduction à l'ancien Français. Paris.
Roby, David Brian (2009): Aspect and the Categorization of States. The Case of ser and estar in
Spanish. Amsterdam.
Rouveret, Alain (1996): Bod in the present tense and in other tenses. In: Borsley / Roberts, (eds.):
125-170.
Schrijver, Peter (2011): Old British. In: Ternes (ed.) (2011):1-84.
Schrijver, Peter (2011a): Middle and Early Modern Breton. In: Ternes (ed.): 359-429.
Schumacher, Stefan (2004). Die keltischen Primärverben. Ein vergleichendes, etymologisches und
morphologisches Lexikon. Unter Mitarb. von Britta Schulze-Thulin und Caroline aan de Wiel.
Innsbruck.
Schumacher, Stefan (2009): Lexical and Structural Language-Contact Phenomena Along the
Germano-Celtic Transition Zone. In: Zimmer, Stefan (ed.): Kelten am Rhein. Akten des
dreizehnten Internationalen Keltologiekongresses 23. bis 27. Juli 2007 in Bonn. Zweiter Teil:
Philologie. Sprachen und Literaturen. Mainz: 247-266 (Beihefte der Bonner Jahrbücher 58).
(English version of: Schumacher, Stefan (2007): Die Deutschen und die Nachbarstämme:
Lexikalische und strukturelle Sprachkontaktphänomene entlang der keltisch-germanischen
Überganszone. In Hans Hablitzel & David Stifter (eds.): Johann Kaspar Zeuß im kultur- und
sprachwissenschaftlichen Kontext (19. bis 21. Jahrhundert), Kronach 21.7.–23. 7. 2006: 167
207. Vienna. (= Keltische Forschungen 2).
Schumacher, Stefan (2011): Mittel- und Frühneukymrisch. In: Ternes (ed.): 85-235.
Stassen, Leon (1997): Intransitive predication. Oxford.
Stassen, Leon (2005): 119. Nominal and locational predication. In: Haspelmath, Martin / Kuteva,
Tania: The World Atlas of Language Structures. Oxford: 482-484. http://wals.info/chapter/119.
Stengaard, Birte (1991): Vida y muerte de un campo semántico. Un estudio de la evolución
semántica de los verbos latinos stare, sedere e iacere del latín al romance del s. XIII. Tübingen.
Ströbel, Liane (2010): Auf den Spuren der Entstehung einer neuen Kategorie. Leerverben als
paralleler Kopulastrang. Frankfurt am Main.
SUPPLETION AND THE COPULA 47
Ternes, Elmar (ed.) (2011): Brythonic Celtic – Britannisches Keltisch. From Medieval British to
Modern Breton. Bremen.
Trudgill, Peter (2011): A Tale of Two Copulas. Language-Contact Speculations on First-Millennium
England. In: Schulte, Michael / Nedoma, Robert (eds): Language and Literacy in Early
Scandinavia and Beyond. Odense and Volda: 285-320 (= NOWELE 62/63).
Trask, Robert L. (1995): Origin and Relatives of the Basque Language. Review of the Evidence. In:
Hualde / Lakarra / Trask (eds.) (1995): 65-97.
Trask, Robert L. (1997): The History of Basque. London.
Trask, Robert L. (2003): Morphology. In: Hualde / Ortiz de Urbina (eds.) (2003): 113-186.
Trask, Robert L. (2008): Etymological Dictionary of Basque. Edited for web publication by Max W.
Wheeler. University of Sussex.
Udolph, Jürgen (ed.) (2013): Europa Vasconica – Europa Semitica? Kritische Beiträge zur Frage
nach dem baskischen und semitischen Substrat in Europa. Hamburg.
Väänänen, Veikko (1981): Introduction au latin vulgaire. 3. éd. rev. et augm. Paris.
Vañó-Cerdá, Antonio (1982): Ser y estar y adjetivos. Un estudio sincrónico y diacrónico. Tübingen.
Vañó-Cerdá, Antonio (1998): Das Schwanken im Gebrauch der Verben ésser und estar im
Altkatalanischen. Iberoromania 47: 27-46.
Vennemann, Theo (2010): Contact and Prehistory: The Indo-European North-West. In: Hickey
(ed.): 380-405.
Villar, Francisco / Prósper, Blanca María (2005): Vascos, celtas e indo-europeos. Genes y lenguas.
[Salamanca].
Wagner, Heinrich (1959): Das Verbum in den Sprachen der Britischen Inseln. Ein Beitrag zur
geographischen Typologie des Verbums. Tübingen.
Wandruszka, Mario (1965): Stare in den romanischen Sprachen. Zeitschrift für romanische
Philologie 81: 423-40.
Wigger, Arndt (2003): Keltische Sprachen: Irisch. In: Roelcke, Thorsten: Variationstypologie. Ein
sprachtypologisches Handbuch der europäischen Sprachen in Geschichte und Gegenwart =
Variation typology. Berlin / New York: 251-276.
Zabala, Igone (2010): Nominal Predication: Copulative Sentences and Secondary Predication. In:
Hualde / Ortiz de Urbina (eds.): 426-448.
Britta Irslinger
Englisches Seminar
Albert-Ludwigs-Universität
D-79085 Freiburg
Britta.Irslinger@mail.uni-freiburg.de
