Causality and Subjectivity: The Causal Connectives of Modern Greek Eliza Kitis

Chapter 8
Causality and subjectivity: The causal connectives

of Modern Greek
Eliza Kitis
1. Introduction
This paper deals with the main (Modern Greek) MG causal conjunctions:
epeii, iati and ioti. There are many connectives in Greek which, besides
their other (main) function or meaning, have either causal connotations or
secondarily are used as causal connectives. The prototypical, however,
purely causal subordinating-connectives in Greek are the ones examined
here. What is interesting to note is that these three causal subordinators are
almost invariably translated as because in English.1 The last one, ioti, has
in the past been considered the high version of the causal connective iati,
and since the climate at the time (after the fall of the junta [1974] and when
demotiki, the low variety, became the official language of the state) was
unfavourable to lexical items originating from katharevousa, the high variety, ioti was not even included in grammar textbooks as a causal connective. However, it is widely used both in spoken but primarily in written
Greek. As both iati and ioti have identical etymologies, they also serve
similar functions as I have shown in previous studies (Kitis 1994, 1996).
On occasion, therefore, I may refer to these two connectives in the singular.
In the present study I want to focus on degrees or types of causality
designated by these connectives, the interrelation between causality and
temporality as exhibited by epeii, and issues of subjectivity relating to the
use of ioti and iati. It will be shown that the functions of these connectives correlate closely with their etymologies and their synchronic meanings are informed by their histories. Their distinct distribution, too, seems
to be a reflex of their historical evolution. This study is part of a larger
project aiming to prove that conceptual domains such as temporality and
causality, as evidenced by Greek connectives, rather than being characterized by discreteness, merge in intricate ways in the use of these connectives. As has been noted, distinct uses of English because-clauses are invariably translated by either epeii or iati and ioti. If there is consistency
226
Eliza Kitis
in modes of translation, this is expected to inform current theories such as

Sweetsers (1990).2 Before addressing these issues, however, let me
summarize some of my previous findings, as they are pivotal in what follows. For reasons of clear exposition I will label epeii as because1(bc1),
on the one hand, and iati and ioti as because2(bc2), on the other. The
examples have been drawn from a large corpus of both conversational and
written language. All examples are real data unless otherwise stated.
Causal connectives have been researched widely in Dutch, French and
German (see Pit 2003 for references), but not in English (but see Breul
1997; Lagerwerf 1998). Because, with certain exceptions, is a glaring omission from research in connectives in the relevance or the neo-Gricean literature, for instance (cf. Blakemore 2002; Carston 2002; Levinson 2000).
The reason for the interest in causal connectives in these other languages
seems to be the various linguistic forms that exist in them, while in English
because (except for infrequent for) is the typical causal subordinator. In
Greek linguistics, apart from my work, there has been scant interest in the
causal connectives of MG (Kalokerinos 1999, 2004), but in Ancient Greek
(AG) there is a study of causal conjunctions that may inform our accounts
(Rijksbaron 1976).
2. The descriptive account

2.1. Distribution of causal connectives
As I have shown elsewhere (Kitis 1994, 1996), only epeii(bc1) can be
preposed (cataphoric) in because p, q structures.3 Whereas epeii(bc1) can
occur either initially (cataphoric or forward) in because p, q or finally (anaphoric or backward) in q because p structures, iati and ioti (because2)
cannot appear initially in them (*because p, q), but always follow their
main clauses (q, because p). In other words, they are always anaphoric.
However, these latter causal connectives can occur, and do occur predominantly, both in spoken (after a longer pause) and in written language (after
a full-stop) sentence-initially, or even paragraph-initially,4 but always following their main complementing it. This is a most prevalent use of these
two connectives. The former subordinator, epeii(bc1), when it appears
finally, is always closely connected to its main, and if it appears sentenceinitially, outside the because p, q structure, the adverbial clause will constitute another speakers turn completing, as in a question-answer adja-
Causality and subjectivity
227
cency pair, its main which is a question; that is, it retains the structure, q
because p:
(1a) iati en ire
o Stefanos?
Why not came-3SG the Stephen?
Why didnt Stephen come?
(1b) epeii (iati) ine arostos.
Because1,2 is ill-MS.
Because he is ill.
(fabricated, both causals possible)
The distribution of these connectives can be summarized pictorially in figure 1:
i. because p, q
epeii(bc1)
*iati/*ioti(bc2)
ii q(,) because p,
epeii(bc1)
iati/ioti(bc2)
iii
q. Because p,
*epeii(bc1)
iati/ioti(bc2)
Figure 1. The distribution of epeii(bc1) and iati(bc2)
2.2. Initial placement of epeii(bc1)-clauses in spoken Greek

It is interesting to note that truly initial placement of because-clauses, preceding the main clause, is not encountered in Schiffrins (1985, 1987) con-
228
Eliza Kitis
versational data. Ford (1993: 17) also reports that in her conversational
corpus, because-clauses are never placed before the material they modify. This is not true of written English, as is well known. In written English, or in some genres (Ford, 1993: 86), initial placement of becauseclauses is not only possible but multiply practiced. Ford attributes the absence of initial because in conversational genre of talk to the lesser degree
of planning that goes into it (86), but also to the possibility of inferring
cause from sequence (89). My explanation, however, differs from Fords
(1993), as it draws on the connectives etymological make-up and its evolutionary meaning (see section 10).
In Greek, on the other hand, initial placement of epeii(bc1)-clauses is
very frequent and absolutely typical of conversational data, too:
(2)
epeii(bc1) I kolitia
sas ine malon
pjo
Because1 the colitis your is rather more

Because your colitis is rather more
ektetameni ap oti sinios
extensive
than what usually
extensive than (what is) expected
kj epeii(bc1) to zaharo
sas
and because1 the blood-sugar your
and because your blood-sugar
en mas epetrepe
na sas osoume kortizoni
not us(acc.) allowed-3SG to you give-1PL steroids
did not allow us to give you steroids,
afto sas
kratisame mesa lio parapano.
for this you(acc.) kept-1PL in a little longer.
thats why we kept you in a bit longer.
(telephone conversation between patient and doctor)
This frequent initial placement of epeii(bc1) in conversational Greek is
considered to be a reflex of aspects of its meaning which derive from its
etymological make-up and its diachronic use (see sections 8 and 10).
229
2.3. Morphological/phonological reduction

The connective iati(bc2) appears in reduced form as ia in some dialects
and in demotic songs (Tzartzanos, V. II: 138):
Strates mou, kaarisete milies mou, fountoite ia a perasi
o ambros.
Paths, clear up, apple trees, send forth leaves, for the bridegroom
will pass.
It can also appear in reduced form as ia in fast speech, just as the English
because, although not so frequently, it can be both phonologically and
morphologically reduced if it is epistemic or performative (as because is
reduced tocause). It is interesting to note that iati(bc2) cannot bear the
main stress in a construction, whereas epeii(bc1) can. The reason for this
potential of epeii as well as for the reduced form of iati will become evident further down.
3. Grounding the discourse

In an earlier paper (Kitis 1996), I argued that epeii(bc1)-introduced preposed clauses, just like initial conditional if-clauses (Ford and Thompson
1986; Haiman 1978), serve a general framework- or background-creating
function for the discourse that follows them (Ford 1993: 14).
The following example comes from an interview with NATOs ex-Secretary General. The interviewer is not quite happy with the answers he got
from Solana regarding his conversion from a staunch anti-Nato campaigner
to NATOs Secretary General, and he decides to rub it in by grounding his
argument in a completely new perspective. This grounding and perspectivization is in this case effected by the use of a very long epeii(bc1)-introduced clause (containing 59 words). Solanas response to it is the second
longest in the specific interview:
(3)
Because1/epeii I come from a country which, like yours, had and

has many prejudices as regards NATO [ etc.] for this (reason)
(thats why) I would also very much like you to tell me [ etc.]
230
Eliza Kitis
The pro-form for this [thats why] substitutes for the topical causal construction. As Haiman (1978: 577) writes left-dislocation apes the discourse situation in which topics are generally established: initial mention in
a full form is followed by subsequent mention in reduced form.
Thematic left-dislocation or a pro-form encapsulating the reason/cause
clause is not a prerequisite for the initial placement of epeii(bc1)-clauses.
Such preposed clauses can preface (un)shared information that serves as
background knowledge for the interpretation of more topical discourse
(Schiffrin 1987: 205), as in the following example, a memo that circulated
at our Department:
(4)
Because1/epeii the loss of the OHP that was in Mr. Xs office will
be reported as a theft, we request that whoever took it or forgot it
lying around should report it at 308A by Wednesday 9/11/04.
Indeed, preposed epeii(because1), just like temporal markers, can function, not only as a topic builder, but also as a topic shift builder and a segmentation marker in continuous discourse (Bestgen and Vonk 2000; Virtanen 1992); but this claim has to be substantiated by a corpus analysis.
In sum, the initial findings of these sections can be stated as follows:
Table 1. Functions of epeii(bc1) clauses
i.
ii.
iii.
iv.
v.
epeii(bc1), but not iati or ioti, can occur in sentence-initial position in

because p, q structures in all genres, including casual conversation.5
Sentence-initial epeii(bc1) is discourse dependent in that its completion (or
part of it) is to be sought in the following discourse.
Preposed because p clauses in because p, q structures have thematization
potential. In this function, the presented proposition is assumed to be given or
undisputed.
On account of the above, and in particular (ii), we can claim that epeii(bc1)
introduces dependent clauses.
On account of the facts summarized in figure 1, it becomes clear that, while
iati/ioti(bc2) can be only anaphoric and it therefore has a backward or retrospective function, epeii(bc1) can be both anaphoric, and hence can have a
backward or retrospective function, but it can also be cataphoric, and hence it
can have a forward or prospective function.
231
It is often stated in the literature that subordinate adverbial clauses are

grouped in language with the logically presuppositional clauses (Givon
1982). This claim, though not explicitly stated, must concern only propositional uses of connectives.6 However, as Givon (1982) stresses, becauseclauses are logically presuppositional but they can also pragmatically
rather than logically background information which functions as a topic.
Although he does not restrict the topic-function potential of becauseclauses to initial placement, that is, to forward causal because, this point is
implicit in his examples. He (1982: 102) concludes: Logical presupposition, involving truth-values, is thus a more limited phenomenon, often
corresponding to but never identical with extreme cases of pragmatic
backgrounded-ness.
It is worth stressing, then, that epeii(bc1) preposed adverbial clauses do
not necessarily (re-)introduce presupposed information, but rather
information that needs to be taken on trust, as having been established in
previous discourse, as given or as indisputable. Owing to this connectives
(because1) potential for (back-)grounding the following discourse, I
dubbed epeii(bc1) polyphonic or heteroglossic (Bakhtin 1981); its adverbial clauses can be likened to quotations from previous discourse
(Akatsuka 1986; Van der Auwera 1986).7
4. The factual causality of epeii(bc1)

In this section, I will demonstrate that only epeii (because1) introduces
factual propositions. To this end I will consider Haimans (1978) claim that
because-clauses are asserted rather than presupposed (573). Although he is
not explicit about it, what Haiman has in mind are q because p
constructions rather than because p, q ones. In other words, he examines
backward or retrospective because. Since in MG we have the option
between at least two causal subordinating conjunctions, it is interesting to
see how English because would be each time translated in Greek depending
on whether it introduces presupposed or new information. Let it be noted
that both connectives are acceptable (except in 5a) if the presuppositional
issue is not heeded.
Haiman considers the timeworn sentence:
(5)
Do you beat your wife because you love her?,
232
Eliza Kitis
which is three-way ambiguous: it can question, depending on where the

focus stress occurs, either (i) the causal relation between the two clauses,
or (ii) either of the two clauses:
(5a) Do you beat your wife because you love her?
(5b) Do you beat your wife because you love her?
(5c) Do you beat your wife because you love her?
(stress indicated)
In (5a) the question is whether the causal relation holds between the two
clauses which represent material or information that is regarded as given.
In (5b), on the other hand, the speaker questions the validity of the content
of the causal subordinate clause as being adequate grounds for the given
action of beating (Given that you beat your wife, do you beat her because
you love her?); whereas in (5c) the speaker questions the validity of the
content of the main clause as being the effect of the causal proposition.
(Given that you love your wife, do you beat her because you love her?).
Of course, Haimans concern is with topical constructions, but topical
constructions need to represent given information. In MG now the situation
is quite different, because given causal material is best structured in
epeii(bc1)-clauses rather than in iati(bc2)-clauses, as can be readily
shown in translating Haimans sentences:
(6a)
xtipas
ti ineka sou epeii(bc1)/*iati(bc2) tin aapas?
Beat-2SG the wife your because(bc1)/* (bc2) her love-2SG?
(6b)
xtipas ti ineka sou *epeii /iati

tin aapas?
Beat-2SG the wife your because*(bc1)/ (bc2) her love-2SG?
ti
ineka sou epeii /?iati
tin aapas?
(6c) xtipas
Beat-2SG the wife your because(bc1)/?(bc2) her love-2SG?
(stress indicated)
In (6a), which is the equivalent of (5a), the acceptable causal conjunction is
epeii(bc1), but not iati(bc2), which, moreover quite expectedly, cannot
carry the main stress. The reason for this situation seems to be that the
causal construction must carry presupposed information since what is
questioned is, not the contents of the two propositions presented in the two
clauses, but rather the causal connection between them. As has been
233
claimed, epeii(bc1), but not iati(bc2), can signal presupposed

assumptions. Moreover, since the focus of the question is the causal
relation between the two propositions, epeii(bc1) is the only choice as it is
the exponent of pure or direct causality and can be, unlike iati(bc2),
topicalized or become the focus of cleft-constructions (Kitis 1994). On the
contrary, in (6b), which translates (5b), iati(bc2) sounds better, while
epeii(bc1) seems to be outright unacceptable (unless [6b] is a rhetorical
question or a quotational one echoing someone elses claim); the reason for
its unacceptability is the questioning of the causal clauses proposition.
Similarly, in (6c), which translates (5c), epeii(bc1) is the natural choice,
while iati(bc2) does not seem to presuppose the truth of the causal
clauses proposition. Rather, the speakers choice of iati(bc2) will
probably indicate his/her doubts about the truth of the state of affairs
presented in the causal clause (a sign of subjectification of causality, see
below).8
The claim made regarding the choice of causal connectives in
translating Haimans examples is corroborated if we try to paraphrase them
making explicit each time the presupposed material:
(6bi) eomenou oti xtipas ti ineka sou, ti xtipas *epeii(bc1)/iati(bc2)
tin aapas?
Given that you beat your wife, do you beat her because(bc2) you
love her?
The unacceptability of epeii(bc1) becomes obvious in an alternative
question in which the two alternants are contradictories:
(6bii) Given that you beat your wife, do you beat her because(iati bc2)
you love her or because(iati bc2) you hate her?
Let it be noted that epeii(bc1) in (6bi) is unacceptable inasmuch as the
question is about whether your loving her or not constitutes the grounds for
beating her, which (the beating) is the topic (given), and not just about the
causal relation; (6c) will be paraphrased as follows:
(6ci) eomenou oti aapas ti ineka sou, ti xtipas epeii(bc1)/?jati(bc2)
tin aapas?
234
Eliza Kitis
Given that you love your wife, do you beat her because(bc1) you
love her?
In (6ci) epeii(bc1) is acceptable if the question is about whether you beat her
or not on the grounds that you love her, whereas iati(bc2) is questionable in
this case.
It appears that whenever the proposition of the causal clause is presented as
an undisputed presupposed fact, epeii(bc1), rather than iati(bc2), seems to
be the preferred causal connective. We can conclude that only epeii(bc1), but
not iati(bc2), can bear both the presuppositional (factual) and the focused
causal relation.
5. Syntactic and semantic constraints on causal connectives

In previous studies (Kitis 1994, 1996), I identified both syntactic and
semantic constraints characterizing the use of (because1)-clauses:
epeii(because1)-introduced clauses have potential for topicalization and
can be left-dislocated (see section 3), can be embedded, and are included
within the scope of the negative operator of the main clause. These constraints do not affect iati/ioti(bc2)-introduced clauses. The latter can be
neither topicalized or left-dislocated, nor can they be included within the
scope of the negative operator of the main clause; iati/ioti(bc2)-introduced clauses also resist embedding.
Moreover, these constraints seem to affect, or rather reflect, the type of
relation that holds between the conjoined clauses. As we have seen, the two
causal connectives (because1, because2) are not freely interchangeable,
even in structure (ii) (fig. 1), that is, when they occur in postposed clauses.
In many cases, quite apart from presuppositional aspects of the introduced
clause, the substitution of epeii(bc1) for iati(bc2) is unacceptable:
(7)
anapsan
fotja iati/*epeii vlepo
kapno
Lit-3PL
fire, because2/*bc1 see-1SG
smoke.
They lit fire, (be)cause I see smoke. (fabricated)
235
While epeii(because1) cannot occur in (7), in some other cases its substitution for iati(because2) or ioti(because2) incurs a dramatic change in the
function of the adverbial clause it introduces:
(8)
an en tis
etroes, a
peenes.
If not them ate-2SG would died-2SG.
If you did not eat them, you would die.
en ihe simasia an
i
patatofloues itan vromikes,
ot did matter if
the potatopeel was dirty,
It did not matter if the potato peel was dirty,
iati
en tis eplenan.
because2 not them washed-3PL.
because they did not wash them.
It is interesting to note that in (8) the because-clause as each time it follows a comma or a significant pause occurs as an intonationally separate
unit. And as such this adverbial clause can admit iati or ioti (because2)
but not epeii (bc1). If the latter causal subordinator replaces iati(bc2),
then it has to be intonationally incorporated within the main clause, and the
meaning of the structure q because p is completely changed: The causal
adverbial epeii(bc1)-clause is included within the scope of the negative
operator. So, what is negated is the causal connection obtaining between
the main and the subordinate clause, as can be shown in the following
notation:
(a)
epeii ~ ((p q)
because1 ~ r)
As I have demonstrated here and elsewhere (Kitis 1996), the proposition

of the adverbial epeii(bc1)-clause is considered factual. In the actual
example, (8), however, what is negated is the proposition of the main
clause only, while that of the iati (because2)-clause is excluded from the
negative scope as its function is simply to explain why the peel was dirty. It
focuses on the lexical item dirty:
(b)
iati
~ (p
q) because2 ~ r
236
Eliza Kitis
iati(because2), therefore, seems to introduce in a separate intonational unit

a clause that functions as an explanation, justification or metalinguistic
comment.
In all cases of the use of epeii(bc1) the causal connection is pronounced or strong. In other words, the function of the construction q
epeii(bc1) p is to actually assert the causal connection between p and q.
This is at least partly the reason why an epeii(bc1)-introduced clause can
be included within the scope of the negative operator. This also explains
why thus introduced clauses can be embedded in reporting verbs. What is
embedded or reported is the causal relation between the two propositions of
the clauses (Kitis 1996). So the question that seems to emerge regards the
kind of causal connection that is signified by the two types of causal connective. Or else, if a causal connection is asserted by epeii(bc1), what is
the function of iati/ioti(bc2)? Kitis (1994: 313) writes in this connection:
The causal connection in all cases [of iati(bc2)] has to be inferred because
it is indirect. By indirect I mean that the explanation arrived at on the basis
of what is stated is not subsumed under a regularity or a generalization. On
the face of it, therefore, sentences do not exhibit the same degree of
explanatory coherence as they would if the accession of the explanation did
not require any extra inferential effort.
And Kitis (1996) underlines that only epeii(bc1), but not iati(bc2), can
signify factuality (see Keller 1995), presupposionality (see Chafe 1984) and
causal relations. The claim that the former connective (because1) only signifies a causal relation between the two clauses is also proven by its potential to be embedded and be included within the scope of the negative
operator. That epeii-clauses only connote factuality and presuppositionality is shown by their intolerance of epistemic modalizers suspending the
factual character of the epeii(bc1)-introduced clause, such as modal verbs
and adverbs, while they can be modified by adverbs strengthening the
causal relation (Kitis 1994):
(9)
Ti
erni iati/?epeii malon en tin
aapa
her-ACC beat-3SG bc2/?bc1 rather not her-ACC love-3SG
He beats her, because2 he rather doesnt love her.
(10) Tis
ini
her-ACC give-3SG
ora
presents
237
sinehos
constantly
iati/*epeii prepi(epist.)
na tin
aapa
bc2/*bc1
must[epist]-3SG to her-ACC love
He gives her presents all the time because2 he must(epist.) love
her. (fabricated, epeii(bc1) in 10 will render must deontic)
As the propositional content of iati(bc2)-clauses is not considered factual or presuppositional, but is rather asserted, these clauses have acquired
a more or less autonomous status; as a result iati(bc2) clauses can at times
be regarded as (near-)paratactic rather than subordinate.
In conclusion, iati/ioti(bc2) clauses resist embedding, intonational incorporation with the main and inclusion within the main clauses negative
scope. Moreover, as iati/ioti(bc2) is not the prototypical exponent of a
causal relation between the two clauses, it cannot be topicalized. In short,
iati/ioti(bc2) clauses cannot be considered presupposed or factual and, as
they are independently asserted and their causal connection to the main is
weak, they are not dependent on the nucleus (main) sentence. It follows
then that while epeii(bc1)-clauses must be truth-evaluable within the complex sentence, iati/ioti(bc2) clauses need not affect the truth-conditions of
the conjunction.
6. Sweetsers solution
In this section I will turn to Sweetsers (1990) account as an obvious source
for a solution to our problem. She claims that we can explain the function
of causal connectives in terms of the three domains she identifies, at which
language supposedly functions, of content-world, epistemicity and speechacts. The examples she cites in relation to causal connectives will be best
translated as follows:
(11a)
(11b)
(11c)
(11a)
John came back because he loved her.

John loved her, because he came back.
What are you doing tonight, because theres a good movie on.
o ianis irise epeii(bc1) tin aapouse
238
(11b)
(11c)
Eliza Kitis
o ianis tin aapouse, iati(bc2) irise

ti kanis to vrai, iati(bc2) ehi ena kalo ero
Let it be noted that (11a) will accept iati(bc2), too. However, the contentworld reading is clearly rendered only with epeii(bc1); iati(bc2) signals a
rather subjective (the speakers) interpretation of the situation (see below).
At first sight it appears that epeii(bc1) functions as a content-world
connective whereas iati(bc2) has an epistemic and speech-act use. Indeed,
Sweetser (1990: 82) writes: My final argument for the existence of these
domains is that there are languages whose vocabularies distinguish more
clearly among the domains than is the case in English.
However attractive Sweetsers solution may appear to be, it loses its
explanatory rigor as soon as we apply her trichotomization to real data
(Kitis 1996). The reason for the collapse of the theory is that there can
hardly be any neat trichotomization of domains, as postulated by Sweetser,
especially between the content and the epistemic domains. As was shown
in Kitis (1994, 1996) in many cases, iati(bc2) functions as an attenuated
causal connective introducing a reason explanation or a justification or
even a comment that is very loosely connected with the main clause or with
some lexical item, or it can even have a meta-linguistic or metadiscursive
commentary function. For example, iati(bc2)-clauses are nowadays overwhelmingly used in advertisements and in most cases there is no clear connection between the iati(bc2)-clause and the preceding main one. It usually functions as a comment justifying why we should buy a product as
admonished in the main clause or even in a preceding NP:
(12) Beauty shop,
iati
i
aniksi eli
ananeosi
Beauty shop, because2 the spring
wants change.
Beauty shop, because2 spring calls for change.
The connection between a directive speech act, which is presumably performed in the elliptical NP-main clause, and the adverbial clause hardly
warrants a speech-act interpretation of the function of this connective, as its
proposition does not refer to the relevance or irrelevance of a state of
affairs as causing or impeding the speakers action (Sweetser 1990: 81).
Consider the gloss: ? We are inviting you to come round to the Beauty shop
because spring calls for change.
239
If we concede that the main clause performs the speech act of inviting
(directive), then one of the (rather peripheral) felicity conditions for this
speech act might be the stating of a reason for speakers performing the act
(Searle 1975). However, whereas in speech act theory this reason is incorporated in the speech act performed, either directly (Why dont you be
quiet?), or by embedding the speech act clause (You ought to be more
polite to your mother), or by making it dependent as an object-clause on
the main expressing the reason (It might help if you shut up),9 in our case
we would have to stretch the notion of reason beyond any generally
acceptable, patterned connection between the two clauses.
This peculiarity is by no means characteristic of Greek advertising
only:
(13) Everybody recognizes status.
Because not everybody has it. (American Bank advertisement)
(13) admits iati(bc2), but not epeii(bc1), in its Greek translation.10 If we
read this as a content-world because, which seems the only option available
in this case, then our world will become a very strange place to live in.
Moreover, as we read this advertisement, indeed we feel an authorial
authoritative voice (persona). Is it then an epistemic or speech-act because?
We seem to be on a wild-goose chase.
Another peculiar function of the same connective, which does not
seem amenable to a Sweetserian interpretation, is the following:
(14) Tha psakso
na o mipos eho
kamia fototipia eo,
Will search-1SG to see whether have-1SG any
copy
I
Ill have a look just in case I have a copy,
iati
to iha
persi.
because it-clitic had-1SG last year
because2(iati[bc]2/*epeii[bc1]) had it last year.
(a colleague and I looking for a document; she suggests that she
should have a look in her office, unmarked [falling] intonation is
assumed, for rising intonation, see Kitis 1994)
(14) cannot be a content-world connective (if it were its paraphrase would
run as follows: Because I had it last year, I may have it this year too; but
not: ?Because I had it last year, I ll have a look inside), neither can it be an
epistemic one.11 The because-clause cannot be the premise for the main as
240
Eliza Kitis
conclusion. (Gloss: ?Since/Because I had it last year I can conclude that

Ill have a look inside! The because-clause is fully sufficient as a cause for
the act of concluding Sweetser, 1990: 80). The only viable interpretation
might be the speech-act one; but then the adverbial clause does not justify
the speech act performed (stating), but rather its propositional content, and
therefore it can be better explained as providing a reason explanation (a
non-nomic cause, Itkonen, 1983) for it (for the action which is described as
imminent).
If (14) can be somehow made to fit Sweetsers thesis, consider (15):
(15) ihame,
ilai,
ti
hiroteri
Had-1PL, namely, the worst
We had, that is, the worst view,
ea
view,
iati
i kaliteri ea evlepe
stin aia sofia.
because2 the better view looked-3SG to the St. Sofia
because the best view was over St. Sofia (church).
(writer talking on the radio about his house in Thessaloniki)
epeii(bc1) is totally unacceptable in this context precisely because there is
no causal connection that is being asserted between the two clauses;12
iati(bc2) does little more in (15) than actually conjoin the two clauses in
an argumentative or elaborative manner. The iati(bc2)-clause elaborates
on the main one.
Whatever tests we may apply to (15), we will not detect any causal or
evidential, or reason relation between the two clauses in whatever domain
we look for its interpretation:
(15) a. *The reason for having the worst view was that the best view was
over St. Sofia.
(15) b. *Because the best view was over St. Sofia, we had the worst view.
(15) c. *Since the best view was over St. Sofia, we had the worst view.
(15) d. *I am in a position to say/assert/know/conclude that we had the
worst view, because the best view was over St. Sofia Church.
Such examples that do not fit Sweetsers thesis, even at the propositional
level considered here, are legion in my corpus.
Initially, her thesis seems to make sense in the case of the Greek causal
connectives: epeii(bc1) is the causal connective used for the content-world
241
domain and iati(bc2) is the connective used for the other two domains,
epistemicity and speech acts. However, Sweetsers thesis collapses under
the strain of real data occurrences of the Greek causal connectives.
Although we might initially claim that the identification of the domains is a
way of getting off the ground, the problem with the meaning and function
of iati/ioti(bc2) looms large. Moreover, epeii(bc1) can be used as a
speech act connective in initial placement even when an explicit
performative verb is absent (Kalokerinos 1999; Kitis 1994, 1996). Quite
clearly, we have to look in a different direction for an adequate explanatory
account. Before turning to other sources that might provide an explanatory
account though, it would be useful to state the findings of the above
sections:
7. The findings
To conclude this rather descriptive, section we may summarize the main
findings in the following table:
Table 2. Functions of MG causal connectives
i.
ii.
iii.
iv.
v.
vi.
vii.
viii.
ix.
x.
xi.
epeii(bc1), on the one hand, and iati/ioti(bc2), on the other, are not
freely interchangeable, even in post-posed position.
iati/ioti(bc2) only can occur sentence/paragraph initially, but always
following the main clause.
iati/ioti(bc2) only can be considered near-paratactic connectives.
epeii(bc1) only is factual.
epeii(bc1)-clauses only can be cleft-constructed.
epeii(bc1)-clauses are always subordinate to the main clause.
epeii(bc1)-clauses only can be included within the scope of the negative
operator, can be embedded and intonationally integrated within the main.
epeii(bc1) in q because p structures always asserts a causal relation
between the two clauses.
epeii(bc1) predominantly can assert causal relations that are considered
direct (also see section 9).
epeii(bc1) only can be the focus by bearing the main stress.
epeii(bc1) signals a cause or a reason irrespective of its placement
(initial/final).
242
xii.
Eliza Kitis
iati/ioti(bc2) can signal a cause or a reason, but it predominantly
appears to signal a rather attenuated form of an indirect causal relation or
sometimes it appears to have an argumentative or elaborative function
(also see section 9).
8. The explanatory account

8. 1. The grammaticalization of epeidi(bc1)
While these findings are interesting, we have not explained why we have
two distinct types of causal subordinating connective each being
characterized by distinct formal and functional characteristics. How can we
explain features such as the assertiveness, but non-factuality, of
iati/ioti(bc2), or the factuality and initial placement of epeii(bc1) even in
conversational data?
Tracing back the history of epeii(bc1), we see that this connective was
primarily a temporal conjunction and a causal one already in Homer. But
originally it probably was a deictic expression having a demonstrative-local
meaning (Schwyzer 1939: 659):
- (= dar-auf )
ep-ei (= dar-auf ) epei
13
epei i epeii
There seems to have been a unidirectional course in the evolution of the

meaning of epeidi from the more concrete domains of locality and
temporality to causality:
demonstrative-local temporal causal
Heine, Claudi, and Hnnemeyer (1991: 50) write that spatial concepts are
more basic than other concepts and therefore provide an obvious template
for the latter. As has been amply demonstrated in the literature, meanings
that were initially conveyed as conversational implicatures were later
established as conventional ones and later entrenched as semantic
meanings. It can be surmised that AG epei, which was a temporal
connective, also acquired meanings of causality, initially by conversational
implicature, that were later mutated to stable conventional implicatures and
semantic meaning. Rijksbaron (1976), who claims that epeidi-clauses of
243
AG have to be interpreted along the same lines as epei-clauses (95), reports

causal implications of temporal epei and epeidi (seeing that, now that)
in Herodotus, although he claims that this observation does not necessitate
assignment of causal meanings to the conjunction (76, 93) (cf. Powel 1949
for a different view). Besides, the most frequently used causal connective
of AG was /ar (Rijksbaron 1976: 185).
This evolution of epeiis meaning, then, may explain a great deal of its
characteristics, namely its initial placement, its discourse-grounding
function, its factuality, its signifying a causal relation between the two
clauses, its potential for presuppositionality and the speakers
uncommitted14 status regarding the introduced clause. The epeii-intoduced
clause bears an iconic relation to its main that can also be called a
consequent in because p, q structures (Kitis 1996). Indeed, we can speak
of antecedent-consequent clauses strictly only in the case of epeii(bc1). In
the case of iati/ioti(bc2) the relation can be so loose that these terms do
not seem to be warranted. So broad is the relation between the two clauses
that ioti(bc2) initiated paragraphs seem to be more than frequent in the
genre of argumentative writing (see notes 4, 20). Resonating Langacker
(1991: 425), one could say that epeii(bc1), but not iati/ioti(bc2), marks a
stative relation that designates the spatio-temporal and causal posteriority
of its trajector (a process) following its landmark.
The speakers uncommitted stance regarding the epeii(bc1) adverbial
proposition is due to its iconic reflection of a temporal order:
event A
event B
Causality invokes the concept of time. As soon as we say why something

happened, we invoke the flow of time (Deutsch 1997).15 The propositions
of epeii(bc1)-clauses, whether depicting events or plans or states of
affairs, are regarded as constants in our speech, impervious to any further
deliberations. Even when the epeii(bc1)-clause refers to the future
(example 4), its proposition is considered to be already in existence at its
allotted moment, and as such it can act as the cause of some effect. So
epeii(bc1)s factivity and perfectivity has little to do with surface
grammatical epiphenomena, such as our tense system, but more with a
244
Eliza Kitis
conceptual brand of factivity and perfectivity that presents events, plans, or

even future arrangements and states as a conceptual fait accomplit. In this
sense, therefore, epeii(bc1) has a rather deterministic character.16
epeii(bc1), then, is a causal subordinating connective that, although
well entrenched in the domain of causality, borders also on the domain of
temporality.17 Yet, due to the characteristics we identified we tend to regard
it as the prototypical subordinating causal connective of MG. It appears,
then, that prototypical causality needs to share a fair amount of the pie with
temporality.
It seems that we are now in a position to complete table 1 (section 3),
which concerns primarily epeii(bc1), as follows:
Supplement to Table 1. Further functions of epeii(bc1) clauses
vi. epeii(bc1) p, q structures exhibit the relation of cause. This relation is
conceptual or logical (logical precedence of cause over effect) but it is coded
semantically, too, in the conjunction.
vii. Since the epeii p (because1 p) proposition in because p, q structures is
considered to be an antecedent to the consequent sentence or discourse, it can
be claimed that the linguistic structure of such because p, q constructions (and
consequently of epeii p, q constructions) is iconic of its conceptual structure.
8. 2. The subjectivity of iati and ioti (because2)

Although iati/ioti(bc2)-introduced clauses are not factual but rather
asserted, one might feel that there is a tinge of non-assertibility about their
proposition. This non-assertibility comes through in their asserted status as
the speakers subjective commitment to the proposition expressed. Whereas
the proposition of epeii(bc1)-clauses is looked upon at a distance as
reflecting an objective state of affairs, the proposition that iati/ioti(bc2)
introduces is filtered through the speakers subjective interface. Therefore,
while epeii(bc1)-clauses are regarded as objective and undisputed (often
due to the temporal anteriority of real events in real time or of
logical/conceptual anteriority) and can be taken on trust, iati/ioti(bc2)introduced propositions may be considered subjective since they have not
been grounded in the deictic spatial, temporal dimension via their
conjunction.
245
Tracing back the history of iati/ioti(bc2), we must first note that iati
is the low variety version of ioti. So, these two conjunctives are the two
faces of the same coin, as we have already seen. It is interesting to note that
the interrogative particle why in MG has the same form as the causal iati.18
Again, this form is the version that originated from the AG preposition
/ia (for/by) and the interrogative pronoun /ti (what?). AG ioti,
therefore, answered to interrogative ? /ia ti? ia touto
oti (for that-DEM that-CONJ). Thus, reanalysis of this conjunction
provides evidence for Heine, Claudi, and Hnnemeyers claim that there is
a shift in functions, from hearer-oriented interpersonal function to textual
function: interrogative structures are reanalyzed as subordinate structures
involving the grammaticalization of question markers to markers of clausesubordination.
[ ?] (deshalb weil)
(=CONJ)
We must also note that /oti in AG is not only a conjunction serving as a
non-factive complementizer to assertive verbs, but primarily to mental and
psychological verbs (Monro 2000: 242). In other words, the origin of /oti
leads us onto paths of subjectivity. As a causal conjunction, /oti appears
to function in AG in exactly the same fashion as does iati/ioti(bc2) in
MG.19
All this can explain why iati/ioti(bc2) may function as a subjectivity
marker introducing arguments (which of course are asserted but
subjective). This subjectivity element inherent in ioti(bc2)s evolutionary
history and etymological make-up can account for its use as a speech-act
connective or an epistemic one, since in both cases there is no explicitation
of the inferential premises drawn upon by the speaker, but rather the
interpreter needs to retrieve implicit inferential configurations for its
interpretation. This causal connective is the one to modulate causes,
internalize and mutate them to reasons, and introduce illocutionary forces.
Its argumentative assertiveness is also explained on the grounds of its
subjectivity. What is filtered through the speakers cognitive interface
comes out as an argument or is looked upon as having the force of one. Its
argumentative force also explains its tendency to become a paratactic
connective:
246
Eliza Kitis
subordinate
hypotactic
(Hopper and Traugott 1993)
paratactic
Since the iati/ioti(bc2)-clauses primary function is to assert the speakers

belief or to state his/her argument, the causal relation is accorded secondary
importance. Moreover, the causal connection can be modalized in a number
of ways, since the proposition does not iconically reflect the out-there order
of the real world or necessarily a nomic proposition. Since iati/ioti(bc2)clauses are not objectively or necessarily causal, the conjunction enters a
process of desemanticization, shedding part of its meaning narrowly
described as causal. And as the causal relation is not focused, the
conjunction cannot be focused upon (no-clefting potential). Indeed,
iati/ioti(bc2) may best be described as a broadly functioning
argumentative discourse marker rather than a causal connective.20
Moreover, iati/ioti(bc2), just like another temporal/causal connective
of MG, /afou (since) (cf. Kitis 2000b), exhibits a high degree of
position variability, as shown in the examples below:
(16) ose
mou ta ramata, fev
iati.
Give-IMP to me the letters,
leave-1SG because2
Give me the letters, because2 Im leaving.
(17) A: to
afises
iana to
plino
[to sakaki]?
it-clitic left-2SG to
it-clitic wash-1SG [the jacket]?
a
to
pas
ston
rafti?
Will it-clitic take-2SG to the
tailor?
Did you leave it [the jacket] for me to wash? Will you take it to
the tailor?
B: mpori
Maybe
A: ine
iati skismeno kses [kseris].
Is-3SG because2 torn
know-2SG
Because it is torn, you know.
(18) S
aapao
afou.
You-ACC love-1SG since
But/Because love you.
247
iati, just like afou (since), can occur medially (17), or can be appended at
the end of a clause (main, 16) transforming it into an explanation,
justification or argument, thus communicatively incorporating the clause in
which they occur into the preceding discourse. This behavior of these two
connectives, which is identified initially at the conceptual or propositional
level of their function, can thus be interpreted as evidence for a process of
grammaticalization and desemanticization, whereby they emerge as
discourse particles. Reanalysis of these connectives will show a shift in
both semantic and syntactic categories affecting the status of both the
proposition and the sentence in which they occur. Thus marked (by
appending or interposing iati and afou) utterances seem to have an echoic
status, in that they do not further the discourse in respect of its
informational increment.
All this leads us to regard epeii(bc1) as the prototypical causal
connective, on the one hand, and, on the other, to view ioti(bc2), but
primarily iati(bc2), not only as an expressive marker, but also as a device
(particle) for organizing conversation, since it functions as a semantically
rather empty co-ordinating connective (Kitis 1994). So Traugotts (1989)
schema, is not only applicable in the case of iati, but it would also have to
reflect iatis organizational potential as a discourse marker and as a
conversation-organizing device:21
Propositional textual expressive22 interactional (or conversation
organizational)
(Traugotts schema extended)
While epeii seems to merge the temporal with the causal domains,
iati/ioti(bc2), on the other hand, is a clear case where the supposedly
out-there content-world is caught in the net of the subjective world that
filters everything that comes its way, internalizing external stimuli as
causes and reason explanations, or simply co-ordinating them as barely
relevant. Indeed, iati(bc2) can be said to be a purely relevance
connective (Blakemore 1987; Moeschler 1993), with procedural function,
orientating the hearer to access the clause it introduces as relevant to what
has preceded it, or as relevant to the main directionality of the discourse
(Koutoupis-Kitis 1982). The acute question that emerges is whether in our
linguistic world, which both reflects and is reflected by the extra-linguistic
world, we can have clear-cut cases and neat categorization such as causality
and temporality, epistemicity and factuality, subjectivity and objectivity.
248
Eliza Kitis
If Keller (1995) is right in his claim that we must distinguish only

between epistemic and factual causal connectives,23 then we can assume
that we are left with the onus of having to account for these two latter types
of connectives. Surely, Sweetsers model, neat and tidy as it may be, is not
adequate for explaining their function, as I have shown. What needs to be
taken on board is their evolutionary course, their etymological make-up and
a careful examination of the versatility of their functions which will be
extracted from a big corpus. What is most essential, though, is to blot out a
neat demarcating line between a content-world domain and an epistemic
one, because, as we have seen with respect to MG causal connectives, the
postulation of these domains cannot afford us an adequately explanatory
account of real data occurrences of these connectives.
As I have shown here by way of considering the main MG causal
connectives, cognitive domains such as the factual or objective and the
internal-subjective rub shoulders producing disconcerting friction as do
domains, such as temporality and causality, which cannot be ripped apart.
In the following section I will discuss the issue of objectivity-subjectivity
that is brought to bear on the use of the connectives under discussion. The
notions of subjectivity, speaker involvement and perspective have often
been identified as underlying factors regulating the use and function of
causal connectives (Pander Maat and Degand 2001; Pit 2003; Verhagen
2000, amongst others). However, my notion of subjectivity is quite distinct
as it relates to types of knowledge and reasoning. In what follows I will
elaborate on this issue and will propose a principled (non-discrete)
distinction between the two domains of subjectivity and objectivity.
9. Subjectivity vs. objectivity
I would now like to draw our attention to a well known and well
appreciated fact of more recent linguistic research, that audible and visible
language is only the tip of the iceberg of invisible meaning construction as
Fauconnier (1997: 1) put it so succinctly. The main bulk of the iceberg that
stays underground, or rather underwater, to carry the metaphor further, is
the more or less well organized system of structured background
knowledge in memory on the basis of which both language and reasoning
function.
This background knowledge, as I claimed elsewhere (Kitis 1982, 1987a,
b, 1995, 1999, 2000a), can be represented in frame and script structures
stored in memory (Minsky 1975; Nelson 1996; Schank and Abelson 1977).
249
In my view, the generic informational content (the flesh) of these structures

can be stated primarily, if not exclusively, in the form of the primordial
conditional. It should be recalled that conditionals are assumed to be
incorporated in the formal structure of universal statements, too (For every
x, if x is S, x is P). Both law and non-law statements may be expressed in
this general form. But law statements warrant inferences of the form if a,
which is S, then a is P. The latter inference is an instantiation, or a token,
with specific values of the former general law-like statement, or the type,
which is expressed in variable form.24
Now, because-clauses can be considered to be the reverse of if pthen
q (p q) conditional statements that are structured in thematic chunks in
our encyclopaedic memory. Moreover, they might on occasion be regarded
as instantiations of more or less general (law-like) statements, that is,
statements regarding specific cases: (i) The water boiled because it reached
100 C, (If water reaches 100 C, water boils [generic]).
It does not seem unfathomable to assume a principled, but fuzzy,
division between two types of reasoning, one firmly based on general
background knowledge-and-belief systems as stored in memory (example
[i]), and another based on more individualized subjective quasi-knowledge
and beliefs. If background knowledge-and-belief systems are generally
accepted, they form the backdrop of language use and language
comprehension, as such systems support inferencing processes. This type
of knowledge can be called, rather schematically, objective knowledge.
What the term objective knowledge needs to bring to the fore is the
intersubjective character of this type of knowledge and beliefs that is the
subject matter of semantic and broad encyclopaedic memory. In the French
(post)structuralist tradition intersubjectivity would be construed as a kind
of intertextuality (Barthes 1970; Kristeva 1969), but not all knowledge and
beliefs need to be the product of other texts (Lyons 1979). In KoutoupisKitis (1982) I postulated two types of background knowledge and beliefs:
Standing Background Knowledge and Beliefs (SBKBs), corresponding
mostly to rather intersubjective, objectivized knowledge, and Current
Mutual Contextual Assumptions (CMCAs), which include local knowledge
and rather subjective beliefs (cf. Kitis 2000a; see also Allen 1995 and
Nelson 1996 for similar distinctions). This type of more individualized
quasi-knowledge and beliefs can be called subjective knowledge.25
Further, we can represent these two domains of knowledge and beliefs
as two bipolar extremes on an epistemic continuum. At one extreme there
will be placed what might be called the objective or intersubjective domain
and at its antipode the subjective domain. Indeed, Searle (1995, 1997, 2002,
250
Eliza Kitis
2004) distinguishes between objective and subjective epistemic claims.

These two extremes are the end points of a conceptual epistemic continuum
connecting the knowledge-substrata of our propositions and, more
generally, of reasoning. Figure 2 is a schematic representation of the
proposed cline. The top level of the cline reflects the world level, both
objective and subjective; the bottom level represents the epistemic
continuum between the two types of knowledge that interact (dotted lines)
with their counterparts at the world level (domains), both feeding them and
being fed by them:
cline
objective domain
objective knowledge
subjective domain
subjective quasiknowledge and beliefs
Figure 2. The epistemic subjectivity-objectivity cline
The objective polar extreme of the proposed cline would attract objective knowledge, nomic regularities and common knowledge, all expressed
in the form of law-like conditional statements (Kitis 1995, 1999, 2000a). In
short, at this extreme there would cluster common knowledge as objective
configurations of subjective knowledge, and, consequently, statements of
direct causality, interpretable as nomic or non-nomic regularities, norms
and internalized rationality principles. Objective knowledge then can be
captured in ifthen conditional clauses. Interpretation of linguistic contributions supported by objective knowledge would also require a more direct
type of reasoning. In short, it would not ordinarily involve inferential leaps
that would require active inferencing procedures on the part of the interpreter. At this end of objective knowledge, one would expect both such
ifthen general statements, but also their instantiations concerning specific
cases. Because-clauses expressing what I called direct causality (Kitis
251
1994) are expected to be converted into such general conditional statements: because p, q or q because p statements would be directly paraphraseable as if p, (then) q26 ones, where both p and q would retain their
propositional form and content, or would be their generalized entailments.27
Consequently, these structures can be instantiated both in epeidi(bc1) and
iati/ioti(bc2) clauses (the latter only in postposed position).
At the other extreme, labeled subjectivity, one would expect a clustering
of non-general knowledge (subjective quasi-knowledge, or beliefs, not
supported by common knowledge); also, uncertain knowledge and individualized configurations of reasoning patterns which have not acquired a
general status and are not, therefore, regarded as objective or factual, or at
times, even acceptable. This type of non-objective (in the sense of nonintersubjective, non-generalized and non-recurring)28 quasi-knowledge and
beliefs cannot be represented in generic p q statements. Intentional causality (Searle 1983) tends to be supported by this latter type of knowledge
structures. Capturing this type of knowledge as conceptually underpinning
the use of causal statements often entails reasoning that requires maximal
inferential leaps on the part of the interpreter; such inferential leaps are not
readily predicted or anticipated on account of systematically stored-inmemory knowledge types. In other words, this proposal subsumes epistemic and speech act uses of connectives under an inferential procedure that
involves inferential leaps, which, however, once explicitated will reveal
conceptual links at the semantic propositional level. On this view then
epistemic and speech act uses of connectives can be regarded as short-circuited conceptual ones. It is this type of short-circuited or gapped reasoning then that would not admit epeidi(bc1) but license iati/ioti(bc2) (cf.
examples 8, 12, 14, 15). Quite apart from speech act and epistemic uses of
causal connectives, this proposal will also encompass many other cases
similar to the ones identified here that do not fall within the one or the other
category. Therefore, this account seems to be more adequate and elegant
since it subsumes under its purview all cases of causal connection.
The suggested cline between the two domains of objectivity and subjectivity (adopting rather general fuzzy terms) may recall Sweetsers (1990)
postulation of the domains of content-world and epistemicity. However,
quite apart from the absence of discreteness in the postulated continuum,
Sweetsers domain of epistemicity is the product of mapping operations
from the sociophysical world, whereas the proposed subjectivity-objectivity
dimension relates to the content of our linguistic contributions, its knowledge-substrata residing in memory, and the brand of reasoning used.
252
Eliza Kitis
Returning to the case of the Greek causal connectives, I would like to

propose, in a rather programmatic fashion, that epeii(bc1) is the causal
connective functioning primarily at the objectivity end, whereas both iati
and ioti(bc2) can handle cases that would cluster around the subjectivity
pole, although they could also occur at or near the objectivity end, especially in fuzzy cases of causality, factuality and presuppositionality. Indeed,
I suggest that this latter connective (iati/ioti(bc2)) should be seen as a
lexical development indexing a domain of subjective individualized quasiknowledge. It can be further proposed that, not only its morphology, but
also its syntactic behaviour (final placement or post-item placement [following the item it qualifies] examples 16, 17) is a development of the conceptual subjective domain from which it derives its function. Subjective,
individualized, non-general quasi-causation is rendered as explanation
rather than as objective causation. The latter (objective causation) need not
be filtered through the subjective interface as it has been amassed and entrenched in public memory. It need not be expanded into an inferential
chain that is assumed but not directly stated in our linguistic contributions.
Objective causation can be linguistically rendered as directly linked clauses
of conditional statements. Moreover, the former (subjective explanation or
justification) ordinarily follows its explicandum, as it relies on subjective
internalized quasi-causation precepts, not readily translated in conditional
statements.
It appears, therefore, that all the characteristics of the two connectives
discussed here fall out of the bipolar continuum between subjectivity-objectivity proposed here. More specifically, epeii(bc1) can be preposed in
an iconic fashion, just as causes in the objective domain (representing the
out-there world) are conceived and perceived as preceding their effects,
while iati/ioti(bc2) always follow their main clause, mirroring our subjectively interpreted causation or reason explanation which ordinarily follows the exlanandum; epeii(bc1)s potential for grounding the discourse,
for factuality and presuppositionality falls out of its firm capacity to frame
propositions in a speaker-detached, so to speak, fashion and present them
as objective and factual states or events, unrelated to the subjectivity of the
speaker.
In sum, what lies behind the notion of a content or propositional causal
connective (primarily epeii(bc1)) is our ability to subsume the relation
between the two clauses within a nomic or non-nomic patterned, or less
patterned, regularity. Such a regularity can in principle be symbolized, as
has been noted, in the form of a conditional statement, p q, stored in
253
memory; epeii(bc1), therefore, indexes causally compact, coherent

propositions; epeii(bc1) guarantees, so to speak, causal and explanatory
coherence. Explanatory coherence entails that there are no huge inferential
leaps to be made that will not be readily warranted by broad encyclopaedic
knowledge stored in or mutated into memory. This statement also explains
the occurrence of epeii(bc1) in both speech-act and epistemic cases when
the speech act or the epistemic verb is explicitly articulated. If the speech
act verb or the epistemic predicate is encoded, then there is no need for reasoning by the interpreter that would require inferential leaps. Moreover,
enablement uses of because, which require longer inferential chains, as in
He will come and spend Christmas with us because Im paying for his
flight, where a free ticket enables him, but does not cause him, to spend
Christmas with us, are rendered with iati/ioti(bc2), but not with
epeii(bc1). The reason for this seems to be the gapped causal inferential
chain that is necessarily generated by the propositions encoded in the two
linked clauses and their entailments.29
Now, if we take on board Kellers suggestion regarding speech-act uses
of connectives, then we can identify the main problem in the data we
examined at the intersection of what Sweetser calls content-world and
epistemic domains. In this section, I suggested that a division of labour
between an objective and a subjective world as a basis for an initial account
of the function of causal connectives might be a better solution than
Sweetsers domains. The transition from norms to nomic p q regularities, to non-nomic regularities, to more individualistic, interiorized and internalized inferencing patterns interpreted as explanation or justification
does not warrant its theoretical translation into the Sweetserian postulated
discrete domains of objective and epistemic worlds. Moreover, we have
seen that the postulation of a subjectivity-objectivity cline does not only
account for, but also seems to be necessitated by, the distinct functions of
causal connectives in Greek.
In the next section, I will compare English because with each one of the
two Greek connectives. The findings are expected to explain some of the
facts that have been pointed out.
10. Comparison of because with ioti (bc2) and epeii (bc1)

Whereas because in English has not been reported to occur in initial position in because p, q structures in conversational data, epeii(bc1), as I have
254
Eliza Kitis
already noted, occurs very frequently in this position in the genre of conversation. On the contrary, just like because, iati/ioti(bc2) does not
occupy initial position in a cataphoric relation to the main clause either in
spoken or written language. That is, it does not occur in because p, q
structures.30 I would like to propose now that this similarity in distribution
between ioti(bc2) and the English causal connective because (in the genre
of conversation) reflects a similarity in function, which in its turn is
explained by the similar etymological make-up of the two connectives.
Because originates from the preposition, by, and the substantive, cause;
originally it was used as a phrase, by cause, followed by the cause or purpose expressed by a substantive governed by of or a subordinate clause introduced by that or why (OED). According to the same source, such subordinate clauses fell into two classes: (a) they expressed cause or reason, and
(b) they expressed purpose. When the subordinate clause signified cause or
reason, that was often omitted, whereas in (b) cases that prevailed in modern usage.
ioti (bc2) (iati being its low variety variant, but both used
interchangeably without significant stylistic variation in MG) originates,
just like because, from the conjunction of the preposition ia(=by) and the
neuter of the demonstrative pronoun in the accusative, touto, which takes
the place of a substantive, followed by the conjunction oti, the equivalent
of that, nowadays called the complementizer:
ia + touto + oti > ia + oti > ioti
Whereas in the case of ioti, the substantive touto gets omitted, in the case
of because it is that that is left out in cause or reason clauses:
by+cause + that > by+cause > because
epeii (bc1), on the other hand, originates form the compounding of the
preposition epi/, (meaning [up]on), and the doric ei/, a locative
adverb meaning where (this point, however, is not confirmed or discussed
in the literature); i/ is an emphatic particle which was occasionally
appended for emphasis and later lost its emphatic function and became a
constant part of the conjunction:
epi + ei > epei, epei i > epeii
255
Whereas epeii grounds its reason or cause as a temporal anteriority in its

spatial anchoring, this grounding is non-existent in both ioti and because;
rather these conjunctions have been the product of a substantive (English:
cause, Greek: touto) introducing a cause or purpose clause (English:
that/of, Greek: oti) by way of a preposition (English: by, Greek: ia). And
thus introduced the propositions of these clauses were probably placed
closer to the sphere of teleology than that of factivity;31 due to their
teleological character then, such because or iati/ioti structures would
ordinarily be postposed, since they signified final cause, purpose or design
(teleology<teleologia<telos[=end, purpose, final cause]). Moreover, it is
worth recalling that language is egocentric and one departs in ones speech
from the occurring or surrounding event or state of affairs to further explore
its possible causes, motivation, explanations or justifications (see Kitis
1996). In other words, in our speech we tend to move from the here and
now of effects to causes, from the explanandum to the explanans. Talmy
(2000) also points to the priority of the resulting over the causing event.
I believe that I have provided, not only an explication of iati/iotis
final occurrence, but also a better explanation for the absence of initial
placement of because in the genre of conversation than Fords (1993)
argument regarding the lesser degree of planning that goes into it (86).
11. Conclusion
In this study I have claimed that there are two types of causal connectives
in Modern Greek, both of which are viewed as purely causal in mass linguistic consciousness, but each one of them reflecting distinct conceptual
and discoursal relations. On the one hand, epeii(bc1) is the prototypical
causal connective par excellence registering cause and effect or event and
explanation relations. The function and meaning of this connective, as I
have demonstrated here and elsewhere, borders on the domain of temporality;32 traces of evolutionary aspects of its meaning as a temporal connective
are still very much evident in its use and function in MG. This comes as no
surprise if one considers that the prototypical causal relation involves two
events both spatially and temporally ordered: event A, which is the cause of
event B, is temporally prior to event B (Mackie 1975). Therefore, the concept of time is inherent in the concept of causal relations. Moreover, the
concept of space is also inherent in both domains of temporality and causality, since events either temporally ordered or causally related are also
spatially anchored.
256
Eliza Kitis
epeii(bc1) is a truly subordinating connective functioning at sentenceinternal level, as the adverbial clause it introduces is closely connected with
its main. Preposed, in because p, q structures, it introduces a clause whose
proposition is considered assumed, undisputed and factual. In short, it is
taken on trust, or as Givon (1982) would put it, it is information that is
shielded from challenge. In this position, epeii(bc1)-clauses seem to
have an iconic relation to their conceptual anteriority. This function of
epeii(bc1) can be traced back to its originally temporal meaning in AG.
When final in q because p structures, epeii(bc1) asserts a cause or a reason
holding between the two clauses. The relation of the adverbial clause to the
main is always causal and its proposition is always considered factual.
However, as the causal relation between the two clauses is asserted,
epeii(bc1) can bear the main stress, can be topicalized, can be intonationally incorporated and embedded within the main clause, is included within
the negative scope of the main and, in short, epeii(bc1)-clauses have all
the characteristics of subordinate clauses. In all these respects, epeii(bc1)
seems to be the equivalent of omdat (Dutch), parce que (French) and weil
(German).
For these reasons, I consider epeii(bc1) the prototypical subordinating
causal connective. The notion of prototypicality of causality as evinced in
connectives, however, carries the implication that the causal connective
displays its prototypical function at the conceptual level. Indeed,
epeii(bc1) can be regarded as the prototypical causal connective functioning at the propositional level. Likewise, Pit (2003: 155) considers
omdat, parce que and weil to be the genuine subordinating causal connectives. Moreover, it is assumed that since it is primarily this type of causal
connective that carries causal conceptual meaning and gives rise to causal
conceptual representations that are brought to consciousness, epeii(bc1) is
also truth conditional, if the primary bearers of truth conditions are not
utterances but conceptual representations (Wilson and Sperber 1993: 23).
iati/ioti(bc2), on the other hand, is the causal connective that functions
at both sentence-internal level (and then it behaves very much like
epeii(bc1)), but primarily at the broader textual and discourse level.
Whereas epeii(bc1) could be characterized as a purely semantic or content
conjunction, since it is the prototypical exponent of causality,
iati/ioti(bc2) plays a primarily cohesive role in discourse, indexing
relevance relations between adjacent, but also remote, clauses, and
257
accessing subjectively tinged causal relations (cf. examples 12, 15). Rather
than bearing conceptual meaning, its role is procedural (Blakemore 1987).
In other words, it can be claimed that only iati/ioti(bc2) can have
procedural meaning, can be more semantically opaque and its meaning
cannot often be brought to consciousness as conceptually represented
(Wilson and Sperber 1993). This broad, but desemanticized, meaning of
iati(bc2) may occasionally allow it to be phonologically reduced in fast
speech. Associated with this broad function is also iatis frequent
occurrence in conversational data as compared to other connectives (also
concerning written language see note 20).
Moreover, iati(bc2), just like afou(since), (but not ioti(bc2), which
occurs more frequently in written language), behaves like a free mover: it
can be interposed or appended to a clause as an interjection or an
expressive marker, and in this function it is used, not only as a relevance
marker, but also as a presuppositional marker accessing supposedly known
information or rather converting the propositional content of the clause to
which it is appended or in which it is interposed (examples 16, 17) to
shared information. Therefore, iati(bc2)s behaviour provides evidence for
the proposed unidirectional grammaticalization process from propositional
meanings to expressive ones (Traugott 1989). However, as iati(bc2) is also
used as a semantically rather empty marker in organizing conversational
moves (cf. example 17, Kitis 1994), we can extend Traugotts schema to
include the interactional level at which currently this connective can also
function.
In sum, while epeii(bc1) can be said to be the causal connective at the
ideational level, iati/ioti(bc2) are the connectives that function at both the
ideational level, but also at the textual and interactional levels of language,
and as a result they are both used in a (near-)paratactic way.33 Its strong
subjectivity element originates from its evolution: it answered to ia ti?
(either explicit or implicit) questions, such questions often involving a
purpose perspective looking into the sphere of teleology (Kalokerinos
1999) rather than factivity and perfectivity as in the case of epeii(bc1),
which never answered to ? (= for what[reason]) questions in AG.
Having said all this, I certainly do not want to claim that these connectives are not interchangeable in a majority of cases. Nuances and meaning
differentials are very subtle in many cases and especially at the propositional or ideational level at which interchangeability seems to be more feasible. These nuances of meaning relate to issues of factuality, generality
258
Eliza Kitis
patterns and issues of objectivity and subjectivity as determined by background knowledge and inferential processes deployed for its accession;
epeii(bc1), then, can be regarded as a formal or material trace for accessing compactly linked causal statements drawn on premises stored in shared
common memory (objectivity marker), whereas iati/ioti(bc2) can be seen
as a formal trace for loosely or idiosyncratically linked statements or for
accessing an inferential procedure that would supply the premises supporting the connection (subjectivity marker). While as far as epeii(bc1) is concerned we can quite confidently conclude that its use marks compact causality drawn on conditional statements stored in memory, iati/ioti(bc2),
on the other hand, can assume both epeii(bc1)s function but additionally
it can also act at the subjectivity end as a discourse marker performing a
multiplicity of functions, not just at the conceptual or propositional level of
language, but mostly at the discourse level of language use. In all these
functions iati/ioti(bc2) does not derive its premises supporting its use
from widely shared or objective background knowledge, as is the case with
epeii(bc1). However, just as the objective and the subjective domains cannot be teased apart but rather are intertwined in intricate ways, so too the
functions of the Greek connectives overlap to a degree that up till now has
not allowed or necessitated their careful examination.
Notes
1. The phonetic transcription of epeii would be epii, but I choose an
orthographical one that reflects the Greek orthography for reasons that will
become obvious. Only this causal connective can at times, and only when it is
in initial position (forward or cataphoric), be translated as as or since
(Stavropoulos 1988).
2. As becomes clear from the statement of purpose, in this study I will focus on
causal connectives at a sentential level rather than at the level of discourse.
Therefore, the data that will be examined will be restricted to one speakers
turn and mostly to the equivalent of what is conventionally regarded as a
sentence (or an utterance). However, the meanings and functions of these
connectives at the propositional level are regarded as preempting their
functions at the broader level of discourse, in my view. But this perspective has
to stay outside the scope of this paper and will only be hinted at at some points.
3. We may distinguish two types of causal structure in this respect: (i) because p,
q in which because functions cataphorically and is therefore a forward causal
connective, and (ii) q because p in which because acts anaphorically and is
259
therefore a backward causal connective. The main clause has been variously
called head, main or nucleus whereas the subordinate because-introduced
clause is the causal clause that can be also called the antecedent with the main
as its postcedent.
4. Also see Bakakou-Orphanou and Koutsoulelou-Michou (1999) on paragraphinitial iati and ioti introducing either subordinate or independent causal
clauses in retrospective evaluative paragraphs for rhetorical effects.
5. Tzartzanos (1989: 141) notes a rare occurrence of initial iati(bc2) instead of
epeii(bc1) in demotic songs.
6. By propositional uses I mean cases in which the two sentences are
straightforwardly joined in terms of the propositions expressed, what I call
direct causality, not assuming subjective correlates.
7. It is interesting to note that Degand (1997) calls the Dutch want (evidently the
equivalent of iati/ioti (bc2), also see Sanders et al 1993 in this connection)
polyphonic, probably because, just like the Greek iati/ioti(bc2), it introduces
a more subjective, argumentative or rhetorical perspective. However, I call
epeii(bc1) polyphonic because it introduces a factual perspective which need
not be the speakers. To use a Lacanian term, it introduces the others
perspective, even if it is considered a generally accepted one. The point is that
this perspective need not be hooked on the speaker. In the same vein, Verhagen
(2000) claims that want is a perspective marker anchored in the speakers
mental space, whereas omdat (evidently the equivalent of epeidi) is not, as it
functions within the perspective set up by the main clause. Indeed, epeidi acts
within the one perspective introduced by the main but this need not be the
speakers.
8. What has been claimed with relation to iati(bc2) carries over mutatis mutandis
to ioti.
9. All examples from Searle (1975).
10. On occasion epeii(bc1) might occur in such cases, but then it is an occurrence
by analogy to iati(bc2) and is considered a rather marked provocative use.
11. If we allow this example to be categorized as a content-world connective, then
the distinction Sweetser wants to preserve between the out-there outside
world and the speakers internal world is completely lost. If, on the other hand,
trying to win the argument, Sweetser would say that on a content-world
interpretation the gloss goes like this: Because I had it last year, this motivates
my looking for it in my office, then, I am afraid, the motivation in this case goes
through my knowledge/conjecture that I might have it. So might it be an
epistemic connective? The circularity of the argument is obvious. (See Kitis,
ms, for criticism of Sweetsers postulation of three domains).
12. It is funny, however, that, although (15) is a typical use of iati(bc2), if you rip
this sentence from its context and ask around for untutored intuitions, people (a
260
Eliza Kitis
fair percentage at least) will tell you that (15) is not correct because they cant
see the connection between the two clauses. The question they ask is: whats
the connection? This means that iati(bc2) is registered in our linguistic
consciousness as a causal connective.
13. It is interesting to note that in Greek, and in other languages I believe, locative
adverbs can have causal meanings, too: pano pou (over that/there) (MG); i is
an emphatic, assertive focus particle, and as such it adds to the factuality of
temporal/causal AG epei (epei+i>epeii) (Denniston 1934). Breathing is
omitted in renditions of AG.
14. uncommitted here is meant to be understood as uninvolved. In other words,
the truth of the adverbial proposition is regarded as generally undisputed and,
therefore, as not an object of negotiation. Rijksbaron (1976) reports that the
epeidi-clauses in AG (Herodotus) refer to a well-known fact which makes in
Dutch a rendering with immers better than one with want (189), while he also
notes that sometimes in epei-clauses an appeal is made to knowledge that is
supposed by Herodotus to be present among his public (182).
15. Kant (Critique of Pure Reason) established the connection between the idea of
succession in time and the idea of causality, a connection that seems to have
been venerated in some form even in the theory of relativity.
16. It would perhaps be intriguing to explore the connection between the function
of epeii(bc1) as a causal lexical exponent in futurity and the future conceived
as a determined factivity (cf. Quantum mechanics, but also Husserl and
Sartre).
17. That is why sometimes sentence initial epeii is best translated as as and
occasionally as since in English. However, the meaning is not quite the same.
For example, neither occurs in the immediate scope of negation (*not as/since
or in alternative constructions such as As/since I am a linguist but *not as/since
I know a lot about this subject), but epeii does. See section 10. Dancygier
and Sweetser (2000: 126) write that since and because express positive
epistemic stance; the clauses in their scope are seen as representing factual or
presupposed information.
18. This interrogative particle can sometimes be used as an interrogative-causal
particle in reduced phonological form primarily in rhetorical questions: a e
les tipote? Why dont you say anything, sklavi mou a e xereste?] My
slaves, why dont you rejoice? (Triantafillidis 1941: 400).
19. According to Schwyzer ioti originates from ia + o ti. In AG o ti or oti (as it
is frequently written) is the neuter of ostis, the relative pronoun, which however
is used as an adverb just like ioti in indirect questions (=for what) (L&S). So,
whether the second compound of ioti originates from this relative pronoun or
from oti, the complementizer (L&S is not specific about it), is a moot point.
What is interesting in both cases is that both o ti and oti have causal meanings.
For a detailed account of oti see Monro (2000: 241-245).
261
20. Sidiropoulous (1994) findings provide evidence for this claim: She notes that
this connective is supplied 24.4% (as opposed to single figures of other
connectives occurrences) in explicitating relations in translating news
reporting material from English into Greek.
21. However, this point has to be taken on trust, since its demonstration would need
a thorough examination of conversational data within a conversation-analytic
framework that would lead us astray in this study. What is important is that
iati is currently used as a discourse particle with reduced semantic content
(having all the functions identified in Schleppegrell 1991), whereas epeii
retains its full share of causal meaning at the propositional level, but is also
used at the level of speech acts -when preposed or when the speech act is
encoded (Kalokerinos 1999; Kitis 1994, 1996)- and explicit epistemicity.
22. Although it is beyond the scope of this paper, I can note that iati(bc2) is
frequently used as a kind of interjection in expressive language as in (i)
(i) As tin efhi sou iati / As to iavolo sou iati (Leave me alone)
(Let you go to your blessing because2 /Let you go to hell, because2
[literal translation]).
23. He claims that the difference between epistemic weil and so-called
illocutionary weil is based on the difference between assertive and nonassertive speech-acts (24).
24. Notice that counterfactuals, too, are supported by such law-like assumed or
presupposed statements stored in encyclopaedic memory (Chisholm 1975, and
more recently Fauconnier 1997, who writes that counterfactuals are not just
fanciful flights of the imagination [14]).
25. The two types of knowledge and beliefs in Kitis (1982) are not identical to the
two types identified here but rather crosscut each other. It must be stressed that
the issue of what constitutes knowledge is a complex one as is the issue of
subjectivity and objectivity, both much discussed in philosophy (cf. Davidson
2001). Here we adopt a rather nave folk notion of what constitutes knowledge
and belief as semantic and episodic memory encountered in cognitive science
literature, roughly corresponding to propositional knowledge that is
independently evaluable (hence objective) in terms of its correspondence to
states of affairs. But as Davidson (2001: 214) stresses, there are no
barriers, logical or epistemic between the varieties of knowledge he
distinguishes that is, between first-person knowledge (of ones own mind),
second person knowledge (of other minds) and world knowledge.
26. See Dancygier and Sweetser (1997) on the intricacies of then in conditional
constructions.
27. This is a rather sweeping generalization and exceptions will have to be
admitted; initial epeii occurrences in what would traditionally be called speech
act cases must be one exception. The view taken here, however, is to regard
262
Eliza Kitis
such epeii occurrences as providing a covert temporal anchor (or even a topic)
rather than just a causal one. This is not the place to go into the inextricability
of causality and temporality, which is at issue here.
28. In his attempt to define causality, Russell points out that events involved in
cause-effect relationships are recurring universals: An event in the statement
of the law [of causality] is obviously intended to be something that is likely to
recur, since otherwise the law becomes trivial. It follows that an event is not a
particular, but some universal of which there may be many instances (Russell
1963: 136). Generally, the prevalent view in philosophy has been that a relation
between two events (or states of affairs) can be said to be causal if it amounts to
a regularity or a causal law.
29. Verstraete (1998) proposes paraphrasing the speech act verb or introducing a
projecting speech act clause that explicitates the internal use (Halliday and
Hasan 1976) of because-clauses. Such explicitation will bring the becauseclause from the level of the grounding and convert it to an Adjunct embedded
in the instantiated type of the main clause. Kitis (1994) had claimed that
epeii(bc1) had the features [+external], [+ideational], [-discursive], [+adjunct],
[+subordinate] and [+direct causation], whereas iati/ioti(bc2) was mainly
characterized as [+internal], [+interpersonal], [+discursive], [+disjunct], [subordinate] and [-direct causality], although both iati and ioti can function
in the place of epeii(bc1), especially in fuzzy cases of presuppositionality, etc.
However, even though Verstraetes solution can take care of both speech act
and epistemic causal connectives, the problem still looms large with cases of
real data as identified here that do not fall into these two categories; these cases
hinge on uncertain knowledge or subjectivized inferential chains that are
neither made linguistically explicit nor supported by compact conditional
statements stored in memory, but rather have to be supplied by the interpreter
as the missing inferential premises of the required reasoning. As claimed, such
premises are not readily drawn from our stored knowledge types. Cf.
Verstraetes examples (i) of external and (ii) of internal conjunction (198):
(i)
(ii)
Bill resigned because his employer wouldnt pay his overtime.

Bill resigned, because his wife told me so.
On my account (i) is readily interpretable on the invocation of a direct

conditional entailed by (i), stored in objective common encyclopaedic memory:
If ones work is not paid one is likely to stop working/resign, etc. The case is
quite different with (ii), which is not readily supported by such a compact
conditional stored as common knowledge in memory: *If one is told by
someones wife that someone resigned then someone resigned. To supply the
inferential process supporting (ii), the interpreter must supply premises that
cannot be stored in common encyclopaedic memory, but are rather drawn on
263
subjective knowledge/beliefs pertaining to specific cases and the roles of the

participants enacted in the speech event. (Gloss would run like this: What s/he
tells me in the main clause bears no direct connection to the because-clause.
(No coherence retrieved). But I know that If someone tells you something then
you are in a position to report it to someone else. Therefore I have to assume
that Bills wife told the speaker what is reported in the main clause and the
because- clause supplies this information as guarantee for the truth of the
proposition of the main clause). (ii) can be rendered only by iati/ioti(bc2). If
(ii) is a clear case of internal conjunction, however, there are many cases that
do not fall clearly within the one or the other category (cf. example 17) but are
rather fuzzy cases on the cline we identified:
(iii)
Shall I turn off the heater? Because2/*bc1 Im airing the room and I
have the window open.
(iii) cannot be a clear speech act case as the because-clause does not just give a
reason for asking, but also a reason for turning the heater off. The inferential
procedure is not straightforward again, but rather involves further premises that
have to be drawn upon to support the causal clause. (Gloss would run like this:
If one airs a room and the window is open the heat of the room is lost. If the
heating system is on under those circumstances then the heat is wasted. If heat
is wasted so is money, etc. Therefore one should turn the heating off to cut
down on the waste, etc.); epeii(bc2) is precluded in (iii) even if the main
clause is declarative (I shall turn off the heater), but is an option if one spells
out the reason a bit further, that is, if one supplies some inferential steps.
30. One might note some exceptions in forward (initial) position in poetic use; in
poetry iati appears to be much more frequent (as a backward [final position]
connective) than epeii, probably due to its subjective character as explained
here (Kitis 1994).
31. Kalokerinos (1999) claims that iati, just like ia na, introducing purpose
clauses, can function in the sphere of teleology, non-factivity and epistemic
causality.
32. It must be stressed, though, that the meaning of this connective is not temporal
in MG.
33. It must be noted that only the low variety-originated iati has been observed to
function at the interactional level, but these functions have not been really
demonstrated in this study. However, both the textual and interactional
functions, which have only been pointed up, follow from iati/ioti(bc2)s
semantic characteristics examined here.
264
Eliza Kitis
References
Akatsuka, Noriko
1986
Conditionals are discourse-bound. In On Conditionals, Elizabeth
Traugott Closs, Alice ter Meulen, Judy Snitzer Reilly, and Charles
A. Ferguson (eds.), 333351. Cambridge: Cambridge University
Press
Allen, James F.
1995
Natural Language Understanding. United States, CA:
Benjamins/Cummings.
Bakakou-Orphanou, Katy, and Stamatia Koutsoulelou-Michou
1999
Evaluation mechanisms in the journalistic text. In Greek Linguistics.
Proceedings of the 3rd International Conference on Greek Linguistics. Amalia Mozer (ed.), 728736. Athens, Ellinika Grammata.
[Bakhtin, Mikhail]
1981
The Dialogic Imagination. M. Holquist, (ed.). Austin: University of
Texas Press.
Barthes, Roland
1970
S/Z. Paris: Seuil.
Bestgen, Yves and Wietske Vonk
2000
Temporal adverbials as segmentation markers in discourse
comprehension. Journal of Memory and Language 42: 7487.
Blakemore, Diane
1987
Semantic Constraints on Relevance. Oxford: Basil Blackwell.
Blakemore, Diane
2002
Relevance and Linguistic Meaning: The Semantics and Pragmatics
of Discourse Markers. Cambridge: Cambridge University Press.
Breul, Carsten
1997
Grammatik und Bedeutung der kausalen Satzverbnde. Because,
as, since und for im schriftsprachlichen Englisch. Tbingen:
Max Niemeyer Verlag.
Carston, Robyn
2002
Thoughts and Utterances: The Pragmatics of Explicit Communication. Oxford: Blackwell.
Chafe, Wallace L.
1984
How people use adverbial clauses. Berkeley Linguistic Society 10:
437449.
Chisholm, Roderick M.
1975
Law statements and counterfactual inference. In Causation and
Conditionals, Ernest Sosa (ed.), 147155. Oxford: Oxford University
Press.
Dancygier, Barbara, and Eve Sweetser
1997
Then in conditional constructions. Cognitive Linguistics 8 (2): 109
136.

2000
265
Constructions with if, since, and because: Causality, epistemic

stance, and clause order. In Cause, Condition, Concession, Contrast,
E. Couper-Kuhlen, and B. Kortmann (eds.), 111142. Berlin:
Mouton de Gruyter.
Davidson, Donald
2001
Subjective, Intersubjective, Objective. Oxford: Clarendon Press.
Degand, Liesbeth
1997
A cross-linguistic view on causation: The case of Dutch and French
causal connectives. Paper presented at the 5th International Cognitive Linguistics Conference. Amsterdam, July, 1997.
Denniston, John Dewar
1934
The Greek Particles. Oxford: Clarendon Press.
Deutsch, David
1997
The Fabric of Reality. England, Allen Lane: The Penguin Press.
Fauconnier, Gilles
1997
Mappings in Thought and Language. Cambridge: Cambridge
University Press.
Ford, Cecilia
1993
Grammar in Interaction. Cambridge: Cambridge University Press.
Ford Cecilia, and Sandra Thompson
1986
Conditionals in discourse: A text-based study from English. In On
Conditionals, Elizabeth Traugott Closs, Alice ter Meulen, Judy
Snitzer Reilly, and Charles A. Ferguson (eds.), 353372. Cambridge:
Cambridge University Press
Givon, Talmy
1982
Logic vs. Pragmatics, with human language as the referee: Toward
an empirically viable epistemology. Journal of Pragmatics, 6 (2):
81133.
Haiman, John
1978
Conditionals are topics. Language 54 (3): 564589.
Halliday, M.A.K.
1994
An Introduction to Functional Grammar, 2d ed. London: Arnold.
Halliday, M.A.K., and Ruqaiyia Hasan
1976
Cohesion in English. London: Longman.
Heine, Bernd, Ulrike Claudi, and Friederike Hnnemeyer
1991
Grammaticalization: A Conceptual Framework. Chicago/London:
The University of Chicago Press.
Hopper, Paul J., and Elizabeth Closs Traugott
1993
Grammaticalization. Cambridge: Cambridge University Press.
Itkonen, Esa
1983
Causality in Linguistic Theory. London/Canberra: Croom Helm,
Bloomington: Indiana University Press.
266
Eliza Kitis
Kalokerinos, Alexis
1999
Etiologies. In Greek Linguistics. Proceedings of the 3rd International Conference on Greek Linguistics. Amalia Mozer (ed.), 141
149. Athens, Ellinika Grammata.
2004
The justifying connection, mostly in Greek. Journal of Greek
Linguistics 5: 2779.
Keller, Rudi
1995
The epistemic weil. In Subjectivity and Subjectification, Dieter Stein,
and Susan Wright (eds.), 1630. Cambridge: Cambridge University
Press.
Kitis, Eliza
1987a
A comment on the article. In Proceedings of the 1st International
Symposium on English and Greek: Description and/or Comparison
of the Two Languages, 4254, School of English, Aristotle University, Thessaloniki.
1987b
A comment on John Hawkins A note on referent identifiability and
co-presence. Journal of Pragmatics 11: 9397.
1987c
Discourse connectors and language learning materials. Journal of
Applied Linguistics 3: 3050.
1994
Some preliminary remarks on the Greek causal conjunctions iati,
epeii and ioti. In Proceedings of 8th International Symposium on
English and Greek: Description and/or Comparison of the two Languages, 306318. School of English, Aristotle University,
Thessaloniki.
1995
Connectives and ideology. Paper read at the 4th International
Symposium on Critical Discourse Analysis: Language, Social Life
and Critical Thought, 1516 December 1995, Athens.
1996
Further remarks on causal connectives of Modern Greek [in Greek].
Studies in Greek Linguistics: Proceedings of 17th Annual Meeting of
the Department of Linguistics, 424436. Aristotle University,
Thessaloniki.
1999
On Relevance again: From philosophy of language across Pragmatics and power to Global Relevance. Journal of Pragmatics 31: 643
667.
2000a
Connectives and frame theory: The case of hypo-textual, antinomial
and. Pragmatics and Cognition 8 (2): 357409.
2000b
Temporality and causality: The case of Modern Greek subordinating
connectives. Revue de Semantique et Pragmatique 8: 121141.
Ms
A reconsideration of some objective dichotomies.
Koutoupis-Kitis, Eliza
1982
Problems connected with the notion of implicature. University of
Warwick Ph.D. thesis.
Kristeva, Julia
1969
Semeiotike. Paris: Seuil.
267
Lagerwerf, Luuk
1998
Causal Connectives Have Presuppositions: Effects on Coherence
and Discourse Structure. The Hague: Holland Academic.
Langacker, Ronald
1991
Foundations of Cognitive Grammar. Vol. II. Stanford: Stanford
University Press.
Levinson, Stephen C.
2000
Presumptive Meanings: The Theory of Generalized Conversational
Implicature. Cambridge, MA: MIT Press.
Lyons, John
1979
Deixis and anaphora. In The Development of Conversation and Discourse, Terry Myers (ed.), 88103. Edinburgh: Edinburgh University
Press.
Liddell, H. G., and R. Scott
1948
Greek-English Lexicon. Oxford: Clarendon Press.
Mackie, J. L.
1975
Causes and conditions. In Causation and Conditionals, Ernest Sosa
(ed.), 1538. Oxford: Oxford University Press.
Minsky, Marvin
1975
A framework for representing knowledge. In The Psychology of
Computer Vision, P. H. Winston (ed.), 211277. NY: McGraw-Hill.
Moeschler, Jacques
1993
Relevance and conversation. Lingua 90: 149-171.
Monro, D. B.
2000
Homeric Grammar. London: Bristol Classical Press.
Nelson, Katherine
1996
Language in Cognitive Development: Emergence of the Mediated
Mind. Cambridge: Cambridge University Press.
OED
Oxford English Dictionary.
Pander Maat, Hank, and Liesbeth Degand
2001
Scaling causal relations and connectives in terms of speaker involvement. Cognitive Linguistics 12: 211-245.
Pit, Mirna
2003
How to Express Yourself with a Causal Connective: Subjectivity and
Causal Connectives in Dutch, German and French. Amsterdam,
New York: Rodopi.
Powel, J. E.
1949
Herodotus. (English transl.). 2 Vols. Oxford: Clarendon Press.
Rijksbaron, A
1976
Temporal and Causal Conjunctions in Ancient Greek. Amsterdam:
Adolf M. Hakkert.
Russell, Bertrand
1963
Mysticism and logic. Great Britain: Unwin Ltd, 1917.
268
Eliza Kitis
Sanders, Ted J. M., Wilbert P. M Spooren, and Leo G. M. Noordman

1993
Coherence relations in a cognitive theory of discourse representation.
Cognitive Linguistics 4 (2): 93133.
Schank, Roger C., and Robert P. Abelson
1977
Scripts, Plans, Goals and Understanding: An Inquiry into Human
Knowledge Structure. Mahwah, NJ: Lawrence Erlbaum.
Schiffrin, Deborah
1985
Multiple constraints on discourse options: a quantitative analysis of
causal sequences. Discourse Processes 8: 281303.
1987
Discourse Markers. Cambridge: Cambridge University Press.
Schleppegrell, Mary J.
1991
Paratactic because. Journal of Pragmatics 16 (4): 323337.
Schwyzer, E.
1939
Griechische Grammatik. Mnchen: C. H. Becksche.
Searle, John R.
1975
Indirect speech acts. In Syntax and Semantics, Speech Acts, Volume
3, Peter Cole, and Jerry Morgan (eds.), 5982. New York/San
Francisco/London: Academic Press.
1983
Intentionality. Cambridge: Cambridge University Press.
1995
The Construction of Social Reality. London: Penguin.
1997
The Mystery of Consciousness. London: Granta Books.
2002
Consciousness and Language. Cambridge: Cambridge University
Press.
2004
Mind: A Brief Introduction. Oxford/New York: Oxford University
Press
Sidiropoulou, Maria
1994
Variation in Translation: English vs. Greek. Athens: Athanasopoulos-Papadamis and Co.
Sosa, Ernest (ed.)
1975
Causation and Conditionals. Oxford: Oxford University Press.
Stavropoulos, D. N.
1998
Reprint. Oxford Greek-English Learners Dictionary. Oxford:
Oxford University Press, 1988.
Sweetser, Eve
1990
From Etymology to Pragmatics. Cambridge: Cambridge University
Press.
Talmy, Leonard
2000
Toward a Cognitive Semantics, Vol. I. Cambridge, MA: MIT Press.
Traugott Closs, Elizabeth
1989
On the rise of epistemic meaning in English: An example of
subjectification in semantic change. Language 65 (1): 3155.
Traugott Closs, Elizabeth, Alice ter Meulen, Judy Snitzer Reilly, and Charles A.
Ferguson (eds.)
1986
On Conditionals. Cambridge: Cambridge University Press.
269
Triantafillidis, Manolis
1941
Neoelliniki Grammatiki [odern Greek Grammar]. Athens: OESB.
Tzartzanos, Achilleas
1989
Reprint. Neoelliniki syntaks. [odern Greek Syntax]. Vol. II.
Thessaloniki: Kyriakid, 1945.
Van der Auwera, Johan
1986
Conditionals and speech acts. In On Conditionals, E. Traugott, Alice
ter Meulen, Judy Snitzer Reilly, and Charles A. Ferguson (eds.),
19725. Cambridge: Cambridge University Press
Verhagen, Arie
2000
Achter het Nederlands. Over interacties tussen taal en
achtergrondcognitie. Inaugural lecture, Leiden University.
Verstraete, Jean-Christophe
1998
A semiotic model for the description of levels in conjunction: External, internal-modal and internal-speech functional. Functions of
Language 5 (2): 179211.
Virtanen, Tuija
1992
Discourse Functions of Adverbial Placement in English. bo: bo
Akademi University Press.
Wilson, Deirdre, and Dan Sperber
1993
Linguistic form and relevance. Lingua 90: 125.

Causality and Subjectivity: The Causal Connectives of Modern Greek Eliza Kitis

Uploaded by

Copyright:

Available Formats

Causality and Subjectivity: The Causal Connectives of Modern Greek Eliza Kitis

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Causality and Subjectivity: The Causal Connectives of Modern Greek Eliza Kitis

Uploaded by

Copyright:

Available Formats

Chapter 8

Causality and subjectivity: The causal connectives

in modes of translation, this is expected to inform current theories such as

2. The descriptive account

Causality and subjectivity

Figure 1. The distribution of epeii(bc1) and iati(bc2)

2.2. Initial placement of epeii(bc1)-clauses in spoken Greek

sas ine malon

Because1 the colitis your is rather more

Causality and subjectivity

2.3. Morphological/phonological reduction

3. Grounding the discourse

Because1/epeii I come from a country which, like yours, had and

epeii(bc1), but not iati or ioti, can occur in sentence-initial position in

Causality and subjectivity

It is often stated in the literature that subordinate adverbial clauses are

4. The factual causality of epeii(bc1)

Do you beat your wife because you love her?,

which is three-way ambiguous: it can question, depending on where the

xtipas ti ineka sou *epeii /iati

Causality and subjectivity

claimed, epeii(bc1), but not iati(bc2), can signal presupposed

5. Syntactic and semantic constraints on causal connectives

Causality and subjectivity

As I have demonstrated here and elsewhere (Kitis 1996), the proposition

iati(because2), therefore, seems to introduce in a separate intonational unit

Causality and subjectivity

John came back because he loved her.

o ianis tin aapouse, iati(bc2) irise

Causality and subjectivity

conclusion. (Gloss: ?Since/Because I had it last year I can conclude that

Causality and subjectivity

8. The explanatory account

There seems to have been a unidirectional course in the evolution of the

Causality and subjectivity

AG have to be interpreted along the same lines as epei-clauses (95), reports

Causality invokes the concept of time. As soon as we say why something

conceptual brand of factivity and perfectivity that presents events, plans, or

8. 2. The subjectivity of iati and ioti (because2)

Causality and subjectivity

Since the iati/ioti(bc2)-clauses primary function is to assert the speakers

Causality and subjectivity

If Keller (1995) is right in his claim that we must distinguish only

Causality and subjectivity

In my view, the generic informational content (the flesh) of these structures

2004) distinguishes between objective and subjective epistemic claims.

subjective quasiknowledge and beliefs

Figure 2. The epistemic subjectivity-objectivity cline

Causality and subjectivity

Returning to the case of the Greek causal connectives, I would like to

Causality and subjectivity

memory; epeii(bc1), therefore, indexes causally compact, coherent

10. Comparison of because with ioti (bc2) and epeii (bc1)

Causality and subjectivity

Whereas epeii grounds its reason or cause as a temporal anteriority in its

Causality and subjectivity

Causality and subjectivity

Causality and subjectivity