Content uploaded by Bernd Heine
Author content
All content in this area was uploaded by Bernd Heine on Nov 06, 2017
Content may be subject to copyright.
Studies in Language : (), –. ./sl...kal
– / - – © John Benjamins Publishing Company
On thetical grammar
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
University of Vienna / University of Cologne / University of Düsseldorf
Most frameworks of linguistic analysis tend to highlight phenomena of language
use and/or language knowledge such as sentence and word structure, while
backgrounding or ignoring other phenomena that are interpreted as being of
more marginal interest for the linguist. e main goal of this paper is to argue
that some phenomena that have previously been treated as being more periph-
eral play an important role in the organization of linguistic discourse, and that
the latter operates in at least two dierent domains, namely that of sentence
grammar and of thetical grammar. Each of the two domains has its own internal
structure, and the two tend to be separated from one another syntactically,
prosodically, and semantically. Building on recent research, the paper aims at
dening the main characteristics of thetical grammar.
Keywords: cooptation, discourse grammar, ellipsis, situation of discourse,
parenthetical, sentence grammar, thetical
. Introduction
In their comprehensive grammar of English, Quirk et al. (1985: 1626) provide the
example in (1) to illustrate the use of commas with adverbials, pointing out that
there is no comma in (1a) while there is one in (1b).
(1) a. He put the chair between you and me.
b. He failed the exam, between you and me.
c. ? Between you and me he put the chair.
d. Between you and me, he failed the exam.
As we argue in this paper, the unit between you and me in (1a) diers from the
corresponding form in (1b) in a fundamental way, in that it is an adverbial in (1a)
but a thetical in (1b).1 is is suggested by the following observations. First, the
unit in (1b), but not that in (1a), is separated by a comma from the rest of the ut-
terance. is dierence has a correlate in the spoken use, where (1b), but not (1a),
On thetical grammar
is likely to be set o by a pause. Second, there is a dierence in placement: e unit
in (1b) is positionally more mobile than that in (1a), as the dierence between (1c)
and (1d) suggests. ird, the unit is an argument of the verb in (1a), while in (1b)
it is syntactically independent from the rest of the utterance. Fourth, there is also
a dierence in meaning and pragmatics: e unit determines and is part of the
sentence meaning in (1a). In (1b), by contrast, the meaning is independent of the
preceding meaning but rather concerns the discourse situation of social interac-
tion. And nally, (1a) can be said to show the literal, locative meaning of between
whereas (1b) is suggestive of a derived, non-locative meaning.2
at theticals such as between you and me in (1b) do not necessarily contrib-
ute to the propositional meaning of an utterance can also be shown with another
example. In a summarizing survey of theticals (parentheticals in his terminology),
Burton-Roberts (2005) presents the English utterance in (2). Rather than to the se-
mantic content of its anchor (or host) utterance,3 the thetical why not have a seat?
relates to the particulars of the situation in which that utterance is made — it adds
in some sense another conceptual “dimension” by turning a one-venue utterance
into a two-venue utterance.
(2) e main point — why not have a seat? — is outlined in the middle
paragraph.
(Burton-Roberts 2005: 180)
e example in (2) also illustrates the following characteristics of theticals:
First, they are typically part of a structure consisting of a thetical plus its anchor,
that is, the utterance where the thetical is interpolated. Second, theticals tend to
be separated by pauses in speech (or commas, fullstops, brackets, and dashes in
writing), they can be interpolated in an utterance without being licensed by any
rule of canonical sentence grammar (henceforth: SG). Nevertheless, they are not
considered by speakers to be “ungrammatical”. ird, they are positionally mobile:
e unit why not have a seat? could be inserted in some other positions of the ut-
terance. And fourth, they can be inserted both in main clauses and in subordinate
clauses. us, in (3), there is one appearing within a relative clause:
(3) For those of us who remember nineteen sixty-ve one or two of our listeners
may Tory party uh leadership contests used to be uh as the cardinals in Rome
and leaders would emerge. (ICE-GB: s1b-024, 1; Dehé 2007: 267)
While not licensed by canonical rules of SG, their internal structure rests on
principles of SG: e thetical has the structure of a prepositional phrase in (1), of
an interrogative clause in (2), and of a declarative clause in (3). Quite commonly,
however, there are some features that are not in accordance with rules of SG. e
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
theticals in (2) and (3) in fact show such features, in that they are elliptic in some
way; we will return to this issue in Section 2.5.
eticals also have in common with utterances formed exclusively by means
of rules of SG that they can have a complex internal organization, which may even
involve syntactic or semantic embedding, cf. (4). e constructed example in (5)
is one that one might not expect to hear regularly but which, nevertheless, was
found to be acceptable to all native speakers of English that we were able to con-
sult. In this example there is an anchor clause (Our chairman is a liar), with four
interpolated theticals, set o prosodically from one another (signalled by com-
mas). And the interpolated theticals can be argued to be interrelated via semantic
embedding, where I believe and you know somehow qualify the meaning of the
immediately preceding unit.
(4) It’s been a mixture of extreme pleasure I’ve had hundreds of letters from all
sorts of people who have enjoyed the book and considerable irritation because
of being constantly interviewed. (ICE-GB; Dehé and Kavalova 2007b: 3)
(5) Our chairman is, everybody is aware of it, I believe, but don’t quote me, you
know the press, a liar.
e example in (5) also illustrates another characteristic of theticals: While
the anchor clause and the interpolated units are syntactically independent of one
another, they may nevertheless be interrelated, e.g. via crossreferencing: e pro-
noun it of the thetical everybody is aware of it is coreferential with the anchor
clause (see Section 3.2).
While they can have a highly complex internal structure,4 there are other kinds
of theticals that have hardly any, or no internal structure, consisting of a single
word, such as the item say in (6) (Dehé and Kavalova 2007b: 1–2; Dehé 2007: 263).
(6) I’d be far more upset if somebody say scratched one of my records than tore
one of my books. (ICE-GB; Dehé and Kavalova 2007b: 2)
e present paper argues for a separate grammatical domain of theticals, one
which subsumes a variety of information units. What unites them are common
syntactic, prosodic, and functional properties which set them apart from “ordi-
nary”, traditional sentence grammar (SG). Although generally treated as periph-
eral, thetical elements can be shown to play an important role in the organization
of linguistic discourse. e paper therefore argues, rst, that there is a distinct
domain of thetical grammar (henceforth TG) in addition to SG and, second, that
both are part of what we propose to call discourse grammar.
Discourse grammar, as we understand it here, is composed of all the linguistic
resources that are available for constructing spoken or written (or signed) texts.
On thetical grammar
e concern of this paper is exclusively with the domain of TG, and more nar-
rowly with the internal structure of TG. e way this domain relates to SG in par-
ticular and the overall structure of discourse grammar in general is an issue that
will be the subject of a separate study.
e main goal of the present paper thus is to present an outline TG as a dis-
tinct domain of discourse grammar. e evidence for this thesis is on the one hand
negative and concerns the external status of theticals: As we hope to demonstrate
in Section 2.1, theticals can be distinguished syntactically, prosodically, and se-
mantically from surrounding linguistic material. On the other hand there is also
positive evidence: eticals have their own placement behavior (Section 2.4), their
internal form diers from that of corresponding information units of SG (Sec-
tion 2.5), and their meaning is determined essentially by the situation of discourse
rather than by syntactic relations within a sentence (Section 2.3).
Being concerned with the nature and the role that theticals play in discourse
organization, the paper is organized as follows: Section 2 is devoted to a general
discussion of theticals, the way they are dened (2.1) and identied (2.2), what
functions they serve (2.3), the features that distinguish them from corresponding
SG units (2.4, 2.5), and in Section 2.6 a distinction between three main types of
theticals is proposed. e ontological status of theticals is briey looked at in Sec-
tion 3.1. Section 3.2 is devoted to cooptation, an operation whereby theticals are
designed by drawing on the resources available in SG. e framework proposed is
related to alternative analyses in Section 4, where we will deal with the question of
how this framework diers from others that have been proposed and what advan-
tages it oers in describing some discourse structures. Finally, in Section 5 some
conclusions are drawn.
Exemplication in the paper is restricted to English. is is done on the one
hand for practical purposes; on the other hand, it reects a major shortcoming
of the paper, namely its Eurocentric scope. While there exists a thorough data
base on discourse grammar and theticals for languages such as English, German,
Dutch, some Romance and Slavic languages, and Japanese and Korean, there is so
far little information on such phenomena in most other languages of the world.
e extent to which the observations made in this paper apply to languages other
than English, especially to non-European languages, is a problem that is outside
the scope of this paper and requires much further research.
. eticals
eticals, or signicant parts of them, have been referred to with a number of dif-
ferent labels, such as parentheticals, parenthetic adjuncts (Corum 1975), disjuncts
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
(Espinal 1991), interpolations, extra-clausal constituents (Dik 1997), inserts, jux-
taposed elements (Peterson 1999: 237), syntactic non-clausal units (Biber et al.
1999: 1082, 1099), supplements (Huddleston & Pullum 2002), and epistemic ad-
verbial phrases5 (ompson and Mulac 1991a; 1991b; ompson 2002: 143), but
most authors converge on calling them parentheticals. Our decision to break with
this tradition by replacing the term parenthetical by the reduced form thetical6
reects that not all instances of the category are in fact parenthetical in the sense
that they are interpolated in or require an anchor utterance.7 Rather, theticals may
be added at the periphery of an utterance, or they may even form utterances of
their own. It is for this reason that we follow Dik (1997: 379) in distinguishing
parentheticals, which “interrupt a clause”, from other kinds of theticals which can
stand on their own, or precede or follow a clause.8 ese other kinds of theticals
are called by Dik extra-clausal constituents (ECCs), a term that he also uses as
a hypernym for both. Our reason for not adopting this term and instead using
“thetical” generally for both his parenthetical and ECC is that “ECC” implies that
the clause, or more generally SG, enjoys a privileged status vis-à-vis theticals. As
we will argue in Section 3.1, such a claim is debatable; the two are assumed to be-
long to separate domains of discourse grammar.9
While there are a number of typological surveys of English theticals (or par-
entheticals), not all distinguish the same range of types or constructions (Espinal
1991: 726–7; Peterson 1999; Dehé and Kavalova 2007b: 1–4; Kaltenböck 2007: 27–
31; Brinton 2008: 9–10). Table 1 lists some typical examples of theticals which
gure in these surveys in some form or other. Note, however, that the theticals
in Table 1 illustrate only one particular type, which we refer to as constructional
theticals and which was the exclusive concern of most previous research; as we will
see in Section 2.6, there are other types of theticals that need to be distinguished
as well.
Table 1. Some constructional theticals of English (see Knowles 1980; Espinal 1991: 726–
7; Lambrecht 1996; Dixon 2005: 233.; Dehé and Kavalova 2007b: 2–4; Kaltenböck 2007;
Brinton 2008; Dehé 2009a; de Vries 2007, n.d.).
Type of constructional thetical Example
Comment clauses Peter will get married next Sunday, I guess. (Espinal
1991: 726)
Reporting clauses e building will, he promised me, be ready on time. (Dixon
2005: 234)
Tag questions He suered great mental distress didn’t he aer the war. (ICE-
GB; Dehé and Kavalova 2007b: 3)
Le-dislocations (As for) Peter, he no longer is my friend.
On thetical grammar
Table1. (continued)
Type of constructional thetical Example
Aerthoughts (e.g., repairs) Jake, I mean Ian, went home already.
Right-dislocations Why, you have to wonder, do they bother with it? e “peace
process”, that is. (e Economist, 28-8-2010: 36)
Phrasal appositives Jake, our boss, told us to stay.
Appositive (or non-restrictive)
relative clauses
And then you had a theologian talking about the Big Bang <,>
which I thought was brilliant. (ICE-GB; Dehé and Kavalova
2007b: 3)
Insubordinated clauses10 If you could open the window, please.
And-clauses Because on this theory and it’s very deeply held uh good
educational news is by denition inadmissible as evidence.
(DCPSE: DI-I01, #91; Kavalova 2007: 147)
. Denition
eticals consist of a word, a phrase, a clause, or even a chunk that does not form
any syntactic constituent. But they may as well have a more complex structure
involving clause combining, as we saw in Section 1. Manifestations of them have
been dened and analyzed in a number of dierent ways11 (for survey discussions,
see e.g. Corum 1975; Espinal 1991; Peterson 1999; Dik 1997; Biber et al. 1999: 1082;
Wichmann 2001; Grenoble 2004; Kaltenböck 2007; Brinton 2008: 7–14, as well as
the contributions in Dehé and Kavalova 2007a), but most authors converge on
portraying them as “non-syntactic” phenomena that pose a problem for syntactic
analysis (Burton-Roberts 2005). According to the denition proposed here, theti-
cals are information units that contrast with the rest of an utterance in having the
properties listed in (7).
(7) Properties of theticals:
a. ey are syntactically independent.
b. ey are set o prosodically from the rest of an utterance.
c. eir meaning is “non-restrictive”.
d. ey tend to be positionally mobile.
e. eir internal structure is built on principles of SG but can be elliptic.
For reasons to be outlined in the remainder of this section, we take (7) to be
a prototypical denition, rather than one based on discrete categorization: e
more of the properties a given information unit exhibits, the better a member of
the category of theticals that unit is.12
ere is both an external and an internal aspect to the behavior of theticals.
Whereas properties (7c) and (7e) concern their internal structure, all remaining
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
properties relate to their external structure and, as we will see in Section 2.2, it is
the latter that is more diagnostic for distinguishing theticals from SG units. We
will now look at each of the properties in turn.
(7a) strictly concerns the external syntax of theticals: ey are not arguments
of any kind of the utterance in which they occur, and they do not form constitu-
ents with units of SG (see e.g. Peterson 1999: 241.). It would in fact seem that
none of the attempts that have been made to dene theticals as part of SG syntax
have really been successful (see Section 4). (7a) has been pointed out in some form
or other by most authors, who observe that theticals are not licensed by canonical
rules of syntax.
eticals been described as disjunct constituents (Quirk et al. 1985: 613, 631;
Espinal 1991; Grenoble 2004: 1972) or as optional expressions, and they dier
from adjuncts in being syntactically unintegrated or detached from the anchor
clause or any other SG structure (e.g., Fortmann 2005; Howe 2008; Quirk et al.
1985: 853; Biber et al. 1999: 1067; see below). ey have also been described as
being syntactically independent in the sense of McCawley (1982). us, in the fol-
lowing example, the deletion of a verb phrase (VP) is not aected by the thetical of
course — that is, VP deletion operates as if the thetical were not there:
(8) John talked, of course, about politics and Mary did too.
(= Mary talked about politics ≠ Mary talked too ≠ Mary talked, of course,
about politics too). (McCawley 1982: 96)
Property (7b) relates to prosody: eticals are typically set o from their an-
chor by pauses and a distinct intonation contour (but see below). For good rea-
sons, a number of authors consider this to be a central or a dening property
(Nespor and Vogel 1986; Rouchota 1998: 105; see Dehé and Kavalova 2007b: 12–3,
Dehé 2009b for more details and discussion), and this property also constitutes
one of the dening features of the extra-clausal constituents proposed by Dik
(1997, Chapter 17). For Pittner (1995), theticals are dened as being separated
from the rest of the utterance via intonation, and Grenoble (2004: 1972) treats (7b)
as one of the two key formal traits distinguishing theticals, the other trait being
that they are disjunct constituents (cf. (7a)). She adds that the separate intonation
contour is “signaled by a higher pitch at the beginning of the thetical and declina-
tion throughout”.
In fact, there appears to be consensus that prosodic non-integration is a pri-
mary characteristic of theticals, and Burton-Roberts (2005: 180) maintains that all
theticals have in common, observationally, “that they are marked o from their
anchors by some form of punctuation in writing or special intonation contour in
speech” (see also Sar 1986: 672; Haegeman 1988: 250; Bolinger 1989; Homann
1998: 300; D’Avis 2005; Dehé 2007: 262; Dehé 2009b). Potts (2003) proposes a
On thetical grammar
syntactic feature COMMA for theticals with “comma-intonation” to signal that
these units do not contribute to assertive content (see Rouchota 1998: 105; Blake-
more 2006: 1673).
at English theticals are characterized by “comma intonation”, or apposition-
al intonation (Dixon 2005: 233), usually oset by pauses, and involve a break in
the intonation curve of the base sentence, has been demonstrated in a number of
studies. Reporting on ndings made by Kutik et al. (1983), Grenoble observes that
the second pause is longer than the rst, with no correlation between pause dura-
tion and the length of the thetical (see also the discussion in Dehé 2007: 262 and
elsewhere). As early as 1974, Sacks et al. (1974: 722) had argued that basically “any
word can be made into a ‘one-word’ unit-type, via intonation”, even if not every
intonation unit is necessarily a thetical.
Being set o from the rest of the utterance by pauses means in written texts
that pauses are marked typically by commas, full stops, dashes, or parentheses. For
example, in the following example, the le boundary is marked by a full stop and
the right boundary by a comma:13
(9) We don’t do drugs, drink, or use profanity. Instead, we instill morals and
values in our boys by raising them right. (BNC; Fraser 2007: 303)
As important as prosody certainly is, it is not really as reliable or predictable a
feature of theticals as claimed by most authors. Wichmann (2001: 181) therefore
warns that there is a certain risk of circularity in taking it to be the only or the
main criterion for dening parenthesis: eticals do not necessarily form a single,
prosodically dened class (Astruc 2005), and quite a number of them are not pro-
sodically set o from their anchor (Dehé and Kavalova 2007b: 14; Dehé 2007).
is appears to apply on the one hand to frequently used ones (e.g. formulaic
theticals such as the ones discussed in Section 2.6). Some theticals develop into
constructional or formulaic patterns and, as a consequence of this process, lose in
prosodic distinctiveness — to the extent that (7b) may no longer be a distinguish-
ing property. On the other hand, it concerns the relative length of theticals: e
shorter a thetical is, the more likely it is that it will be aected by and/or interfere
with the prosodic structure of its anchor. For example, Dehé (2009b) found that,
unlike types of longer theticals such as sentential parentheticals and non-restric-
tive relative clauses, shorter theticals such as comment clauses, reporting clauses,
vocatives, and question tags do not show a strong tendency to be prosodically
separated. We therefore take (7b) not as necessary but — in conjunction with the
other criteria of (7) — as sucient criteria, that is, the presence of separate in-
tonation contours and/or pauses is indicative of the presence of a thetical while
absence does not mean that the relevant unit is not a thetical.14
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
In response to Dik’s (1997) denition, alluded to above, Tucker (2005: 694) ar-
gues that “adjunctival expressions, when intended as separate information groups,
are commonly marked o prosodically from the rest of the clause of which they
are clear elements”. We surmise that exactly when intended as separate informa-
tion groups and marked o prosodically from the rest of the clause, the expres-
sions concerned are no longer adjuncts but rather theticals. e dierence be-
tween adjunct and thetical can be illustrated by the item really in (10): Whereas the
adjunct really in (10a) is prosodically integrated, the thetical in (10b) is not, as the
comma indicates, and whereas the adjunct is positionally restricted, the thetical
can occur either nally (10b) or initially (10c).
(10) a. Bob is really a poet.
b. Bob is a poet, really!
c. Really, Bob is a poet!
And there is another dierence in example (10), which may be of help in
also describing property (7c), namely what, following Huddleston and Pullum
(2002: 1350–62), we call “restrictive meaning”. ese authors propose distinguish-
ing a construction type that they refer to as supplements. is type, which largely
corresponds to our notion of theticals, diers from integrated constructions, i.e.,
sentence grammar units, in being syntactically and prosodically non-integrated
and semantically non-restrictive. us, the item really in (10a) is an adverb that
is restrictive in the sense that it determines the meaning of the predicate. In (10b)
and (10c), by contrast, it is non-restrictive in that is not semantically part of the
clause Bob is a poet but rather concerns the situation of discourse — in this case
more specically the nature of speaker-hearer interaction. We will return to this
issue in Section 2.3.
According to property (7d), the placement of theticals is mobile (Brinton
2008: 18; Wichmann 2001: 179). In fact, many can occur virtually anywhere with-
in an utterance. Still, there are constraints on placement, most of all relating to
their discourse-specic functions (see Emonds 1973: 338, Peterson 1999: 237–40,
Grenoble 2004: 1966–7, or Brinton 2008: 8 for discussion of English comment
clauses), and certain theticals are restricted to specic positions of an utterance;
we will return to this issue in Section 2.4.
With reference to (7e), Burton-Roberts (2005) concludes that (paren)theti-
cals “range from manifestly non-syntactic phenomena to what are oen regard-
ed as central syntactic constructions”, and Dehé (2009a) characterizes their in-
ternal structure in the following way: “Parentheticals are expressions of varying
length/complexity, function and syntactic category, […]. Expressions that have
been argued to be parenthetical in nature include sentence adverbials and adver-
bial clauses, one-word expressions (e.g., English like, say, what), comment clauses
On thetical grammar
(e.g., English I think, I suppose, you know, German glaube ich, French je pense),
reporting verbs (e.g., English he said, said she), vocatives, nominal appositions,
non-restrictive relative clauses (NRRC), tag questions, and various types of full or
elliptical clauses”.
(7e) raises two central questions in particular, namely (i) what material can or
cannot be ellipsed, and (ii) why is it ellipsed? Question (i) will be taken up in Sec-
tion 2.5, while (ii) can essentially not be answered without recourse to an appro-
priate theory of discourse organization. Such a theory is beyond the scope of the
present paper, which is restricted to dening an outline of some formal properties
of discourse grammar.
In the catalog of properties in (7) we ignored a number of features that are
also mentioned in works on the subject. One of these features concerns the medi-
um of language use: eticals are widely assumed to be a characteristic of speech,
and a number of them are felt to be somewhat unusual when found in writing. It
would seem, however, that theticals are much more common in written texts than
is widely assumed. According to another feature that is frequently mentioned, the
use of theticals is optional or “non-essential” in that they can be dropped without
aecting the grammatical acceptability of an utterance (cf. Dik 1997; Fortmann
2005; Schneider 2007a; 2007b; Howe 2008). We propose not to include this as a
dening property since it does not appear to be suciently discriminative; it also
applies to many constituents of SG, such as adverbial adjuncts or relative clauses
(cf. Tucker 2005: 694). Note further that theticals can occur as independent utter-
ances (cf. 2.2), and one may wonder what “optionality” would mean in such cases.
. Identifying theticals
On the basis of the properties listed in (7) it is possible to propose a few criteria
that are of help in identifying instances of theticals, to distinguish e.g. the thetical
between you and me in (1b) from the corresponding SG unit in (1a). e catalog in
(11) is restricted to some general observations.
(11) Possible diagnostics of theticals:
a. e occurrence of a unit is not licensed by the syntax of its anchor.
b. e unit is set o from its anchor by its intonation contour and/or
pauses in speech or by punctuation marks in writing.
c. e unit shows an extraordinary variability in its placement within the
utterance.
d. e internal structure of the unit is elliptic.
Perhaps the most readily available tool for identication is provided by (11a):
Since thetical categories are autonomous syntactic units, many of which can form
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
utterances of their own, any text piece that is not licensed by canonical SG is po-
tentially suggestive of a thetical category, as angry at the delay in (12a) is.
e examples in (12) also illustrate (11b): Commas, full stops, dashes, and
parentheses in English and many other languages may signal boundaries of theti-
cals. In a number of cases, punctuation provides the only means of identication
in written discourse, while in speech it is distinct intonation contours and pauses
instead.
Since theticals are as a rule much more exible in their word order and their
scope behavior than corresponding SG units (7d), their position within an utter-
ance may also serve as a diagnostic, cf. (11c). For example, when certain English
adverbs occur clause-nally, this is diagnostic of their thetical status, as in (12b),
since as units of SG they are disallowed in this position, cf. (12c).
(12) a. e chairman, angry at the delay, demanded a full report. (Kaltenböck
2007: 30)
b. John has robbed a bank, allegedly/regrettably.
c. * John has robbed a bank allegedly/regrettably. (Stowell 2005, (15))
(11d) concerns a property that will be looked at in Section 2.5, namely that
the internal structure of theticals can be elliptic in a way that corresponding infor-
mation units of SG cannot. Accordingly, if one nds a text piece that lacks some
constituent that would be required by canonical rules of SG then that piece can
be suspected to be a thetical. For example, the English verb know normally takes
a complement, as in (13a), but lacks a complement in (13b), hence you know in
(13b) may be suggestive of a thetical unit — an analysis that is supported by the
presence of a comma (see above).15 Note, however, that theticals frequently are,
but need not be elliptic; accordingly, (11d) works only in one direction — that is,
absence of ellipsis cannot be taken to be indicative of absence of thetical status.
(13) a You know that there’s no money to be made out of recycling.
b. You know, there’s no money to be made out of ri-, recycling.
c. You know there’s no money to be made out of recycling. (Biber et al.
1999: 197)
But there are types of utterances where none of the diagnostics really applies.
One type concerns utterances of the kind illustrated in (13) where the matrix
clause can occur either with or without a complementizer (see also 3.3). ere
may be semantic clues providing further evidence. According to the interpreta-
tion proered by Biber et al. (1999: 197), You know that in (13a) has a SG mean-
ing (‘you are aware that …’), while in (13b), you know is a comment clause, i.e. a
thetical, where the speaker tells the addressee something the latter perhaps does
not know and the function of you know is to underline the truth of the statement.
On thetical grammar
In (13c) nally, both meanings are possible, hence (13c) is ambiguous. Accord-
ingly, the status of you know in (13c) is unclear: It may be either an SG unit without
a complementizer or a thetical.
A review of the relevant literature shows that the status of initial clauses such as
you know in (13) poses not only a problem for our denition in (7) but also for other
researchers (see Kaltenböck 2010: 247). Some scholars take both (13a) and (13c)
to be suggestive of theticals (e.g. Kruisinga 1932: 486; Ross 1973; ompson 2002;
Kärkkäinen 2003), while others treat them as SG units (i.e. matrix clauses; Svensson
1976: 375; Stenström 1995: 293; Peterson 1999: 236), and still others analyze them as
ambiguous, allowing an interpretation as both theticals and SG units depending on
the context and type of matrix predicate (e.g. Urmson 1952: 481; Aijmer 1972: 46;
Quirk et al. 1985: 1113; Biber et al. 1999: 197; Huddleston and Pullum 2002: 896).
e second type concerns utterances in isolation: How can stand-alone theti-
cals be distinguished from ordinary SG units? at the information unit it’s true is
a thetical in (14a) through (14d) is uncontroversial on account of diagnostic (11a).
But what about the status of it’s true in (14e), which can be either a thetical in isola-
tion or an utterance of SG? is question cannot be answered at the present stage
of research as it requires much further work on the prosody and the discourse
status of the information units concerned.
(14) a Kim, it’s true, went to Seoul on Sunday.
b. Kim went, it’s true, to Seoul on Sunday.
c. Kim went to Seoul, it’s true, on Sunday.
d. Kim went to Seoul on Sunday, it’s true.
e. It’s true.
ere is, however, evidence to the eect that theticals can form utterances of their
own, without requiring an anchor. For example, following a number of other au-
thors (see Stowell 2005), we observed in Table 1 that non-restrictive relative claus-
es are constructional theticals. Restrictive relative clauses, by contrast, are gener-
ally classied as units of SG. Now, Burton-Roberts (1999: 37) found that it is only
non-restrictive relative clauses that can form utterances of their own. us, when a
relative clause is added by another speaker, as in the following example, it must be
construed as a non-restrictive and never as a restrictive relative clause:
(15) A: My publications will include the article in Scientic American.
B: Which you’ve not even begun to write yet. (Burton-Roberts 1999: 37)
Nevertheless, the diagnostics of (11) are helpful but do not take account of all
instances of presumable theticals. In accordance with our denition in (7) we will
say that the greater the number of the properties listed that can be identied, the
more likely it is that a given information unit qualies as an instance of a thetical.16
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
. Functions
Quite a number of generalizations have been proposed for the meaning or func-
tions of theticals (e.g., Grenoble 2004; Blakemore 2006; Rouchota 1998; Hud-
dleston and Pullum 2002: 1350–62). eticals have been described as serving to
specify, exemplify, explicate, clarify, characterize, elaborate on, or delimit a ref-
erential unit introduced earlier (Mazeland 2007; Dehé 2009a). And it has been
argued that theticals do not contribute to the assertive content (Potts 2002) or the
truth-conditionality of an utterance (Urmson 1952), that they cannot be ques-
tioned in a standard way (Espinal 1991: 729), that they are impervious to negation
(Espinal 1991: 757; Dixon 2005: 235–6) or tend to be of positive import (Jackend-
o 1972: 97), that their meaning necessarily relates to their anchor (Taglicht 1998),
that they function as a comment17 or gloss on some aspect of meaning of their an-
chor (Espinal 1991: 757; Rouchota 1998: 102), that they serve mitigating functions
(Fraser 1980; Schneider 2007a; 2007b), or have a role in information structuring
(Taglicht 1984: 22–25; Brandt 1996).
Such generalizations are helpful toward understanding salient functions of
theticals, but they do not seem to apply to all of them (see, e.g. Ifantidou-Trouki
1993; Ifantidou 2001; Blakemore 2006: 1684; Dehé and Kavalova 2007b: 9–10), nor
are they all really more characteristic of theticals than of some other kinds of in-
formation units. For example, lack of truth-conditionality does not apply generally
to theticals; evidential theticals, such as I hear or evidently, are truth-conditional
(Ifantidou 1994; Rouchota 1998: 113). And the thetical you say in (16a) appears to
make an essential contribution to truth conditions since the speaker will be un-
derstood to be committed to (16b) but not to (16c). A similar example is provided
by Potts (2002: 661) in (16d), which does not have the truth conditions of George
is a spy.
(16) a. Because John is, you say, a spy, we should be careful what we say to him.
b. We should be careful what we say to John because you say he is a spy.
c. We should be careful what we say to John because he is a spy.
d. George is, you say, a spy (but I don’t believe it). (Ifantidou, 2001: 149;
Blakemore 2006: 1681)
Furthermore, that theticals are impervious to negation is a generalization that
is in need of qualication. It does not apply to instantaneous theticals, cf. (17a),
and constructional theticals can also be negated in appropriate contexts, as in
(17b).18
(17) a Mary — don’t forget — is coming over to visit. (Kaltenböck 2007: 40)
b. I wouldn’t put anything on it I don’t think because two feet isn’t very high.
(ICE-GB: s1b-025 #12; Dehé 2009b: 596)
On thetical grammar
Nevertheless, it would seem possible to propose a kind of lowest common
pragmatic denominator. Grenoble (2004: 1954) argues that parentheticals can be
dened functionally but not structurally because “they do serve clear pragmatic
functions, and are united in operation on a distinct discourse plane”. In her view
they express comments about the utterance, although she concedes that “the def-
inition is vague, as these comments include subjective evaluation, addresses to
the interlocutor, speaker point of view, and so on” (ibid.). In fact, the meaning
of theticals has been described as being metacommunicative or metatextual in
nature (e.g., Bayer 1973; Petola 1983: 103; Pittner 1995; Traugott 1995: 6; Grenoble
2004: 1953); rather than being restricted to and determined by the syntax of the
sentence, it relates to the situation in which discourse takes place. We therefore fol-
low Huddleston and Pullum (2002: 1350–62) in calling this meaning “non-restric-
tive” (see Section 2.1). While this meaning does not surface in all instances of thet-
icals (Pittner 1995), it extends beyond the scope of SG, as is suggested e.g. by the
analysis of information structure (cf. e.g. Brandt 1996 on the focus-background
structure in German), being shaped by the entire situation of discourse. e exact
nature of this situation is a matter of further research; nevertheless, salient char-
acteristics of it can be described with reference to a network of the components
proposed in (18) (cf. Kaltenböck 2010: 263).
(18) Components determining the situation of discourse:
Text organization (TO)
Source of information (SI)
Attitudes of the speaker (AS)
Speaker-hearer interaction (SHI)
Discourse setting (DS)
World knowledge (WK)
Obviously, the components AS and SHI relate, respectively, to the notions of
subjectication and intersubjectication of Traugott (2003); but reducing the dis-
course situation to such concepts would not do justice to the network architecture
characterizing the situation of discourse. In the following we will look at each of
the components in turn.
Text organization (TO). is component relates in particular but not only to what
Blakemore (2006: 1684) calls grammatical parentheticals, which contribute to
what is implicitly or explicitly communicated by the anchor utterance. Qualifying
or modifying the utterance meaning by providing supplementary information for
coherent interpretation is a major function relating to this component, cf. (19a).
Another common function concerns repairs, including self-correction, contras-
tive reformulations, etc., cf. (19b–c).
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
(19) a. He called John — he is one of his best friends — to nd out what had
happened. (Kaltenböck 2007: 29)
b. He is, or at least believes that he is, a major composer. (Quirk et al.
1985: 976)
c. I’ll see you tomorrow morning, I mean, aernoon. (Brinton 2008: 129)
Source of information (SI). Another major function of theticals is to provide the
source of information of an utterance, cf. (20a). It is most of all reporting clauses
that manifest this component, as in (20b), and quite a number of uses of this com-
ponent concern evidential concepts, cf. (20c).19
(20) a. e issue is, as you say, extremely delicate. (Blakemore 2006: 1672)
b. Next year, John announced, I will move to London. (Kaltenböck 2007: 29)
c. e MP, they say, has never been to his constituency.
Attitudes of the speaker (AS). Oering evaluations, modal assessments, opinions,
etc. is another salient function of theticals, cf. (21a). Many comment clauses be-
long here (Brinton 2008), cf. (21b), but there are also various other types of theti-
cals that relate to this component.
(21) a. It is also, perversely enough, an unintended validation of Chomsky’s
disruptive agenda. (T. Givón, Funknet circular, 12-03-2010)
b. You could, I suppose, commission some prints of you yourself.
Speaker-hearer interaction (SHI). is component relates to the particulars of
speaker-hearer relations, as in (22a) and (22b), but may also aim at inducing the
hearer to take some action, as in (22c) and (22d), possibly also with a tag question,
as in (22e).
(22) a. Very short skirt on if you don’t mind me saying. (ICE-GB: s1a-040–089)
b. Mary — I hate to tell you this — is coming over to visit. (Kaltenböck
2007: 39)
c. Mary — don’t forget — is coming over to visit. (Kaltenböck 2007: 40)
d. What is AIDS? Quickly! (Espinal 1991: 755)
e. Always poking her damn nose in isn’t she. (ICE-GB: s1a-007–224)
Discourse setting (DS). eticals associated with this component invoke the extra-
linguistic situation in which the utterance is made. is situation can comprises
the coparticipants and immediate physical surroundings, as in (23a), the manner
of communication, as in (23b), the communicative act itself, as in (23c), or other
events happening during the speech event, cf. (23d). It can also be taken to include
assumed shared knowledge if it can be related to the preceding discourse, as in
(23e).
On thetical grammar
(23) a. So what we can do in fact I’ll just turn it o <,> is to use that signal to
train people’s ability to perceive voicing. (ICE-GB: s2a-056-87)]
b. He failed the exam, between you and me. (Quirk et al. 1985: 1626)
c. And it considered that the exclusive emphasis on the linguistic foundations
<,> of the system had led to the neglect of the computational possibilities
<,> particularly interactive uhm operation of the system <,> I’ll come
back to this aspect later on <,> and that in insucient attention was
paid to uh dictionary compilation. (ICE-GB: s2a-032-59)
d. He designed a torsion apparatus to measure the friction <,> excuse me
<,> and found some most surprising results. (ICE-GB: s2a-041-1)
e. at, of course, means plenty of the aky white stu. (Biber et al.
1999: 773–4)
World knowledge (WK). e speaker may appeal to knowledge beyond that de-
rived from the situation of discourse but presumed to be shared by the interlocu-
tors in justifying the utterance or placing it in a wider context:
(24) a. But we were also deeply interested in structure — how could a linguist
not be? — provided it were studied and explained in its relevant broad
connectivity. (T. Givón, Funknet circular, 12-03-2010)
b. Of course, on Monday nights they settle down to watch — what else —
“Murphy Brown” (Saturday Evening Post, 1992; Lenk 1992: 189–202;
Brinton 2008: 212).
To conclude, theticals, such as between you and me in (1b), dier from corre-
sponding units of SG, as in (1a), in drawing on an additional functional dimension
shaped by the components distinguished in (18). ese components appear to
form a network of relations that is activated by speakers (or writers) to “step out”
of the connes of SG in order to place their communication in a specic situation
of discourse. Rather than being mutually exclusive, the components interact in
shaping the meaning of theticals. Depending on the intentions of the speaker and
the information unit selected, one specic component can be foregrounded while
others are backgrounded. If that component is TO then the resulting meaning of
the thetical is likely to be similar to that of SG units, which also concern the linear
arrangement of information units. But rather than one component, foregrounding
is likely to concern some combination of the components.
Accordingly, the meaning of a thetical can be, and frequently is, complex. For
example, in the utterance of (25), I prefer not to know how awful can be understood
both as relating to TO, modifying the noun paper, and to AS, i.e. the attitudes of
the speaker.
(25) He gave this I prefer not to know how awful paper. (Espinal 1991: 748)
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
In a similar fashion, a number of the examples proposed above to illustrate a
given component are not necessarily restricted to that particular component but
may as well be interpreted with reference to more than one component. For ex-
ample, we just noted that the meaning of the thetical he is one of his best friends in
(19a) is likely to resemble that of a corresponding SG unit and could conveniently
be paraphrased by a relative clause. Nevertheless, there is reason to argue that the
thetical of (19a) is not restricted to TO but also relates to SHI in that it signals
relevance to the current discourse situation and may thereby be instrumental to
ensure smooth processing.
Whether, or to what extent, it is possible on the basis of the form of a thetical
to predict which components are involved in its meaning is a task for future re-
search. Nevertheless, it seems possible to propose a few tentative predictions, such
as the following:
a. Reporting clauses (e.g., he said) are likely to foreground component SI.
b. Comment clauses (e.g., I think) are likely to foreground component AS.
c. Tag questions (e.g., isn’t it?) are likely to foreground component SHI.
d. Appositives (noun phrases or relative clauses) are likely to foreground compo-
nent TO.
. Placement
One main characteristic of theticals is that they tend to be both syntactically and
prosodically autonomous and positionally mobile (see (7), Section 2.1). Many can
be placed essentially anywhere within an SG unit but the opposite does not apply:
eticals form fairly monolithic blocks that as a rule cannot be split (cf. Espinal
1991: 751), at least in English, as the following example illustrates:
(26) a Mary — I just learned from her daughter — is coming over to visit.
b. * Mary — I just learned — is coming over to visit — from her daughter.
c. * I just learned — Mary is coming over to visit — from her daughter.
e extent to which placement of theticals relates to the meaning of an utter-
ance is an issue that requires much further research. What surfaces from the litera-
ture, however, is that placement signicantly aects the semantic-pragmatic scope
relations within an utterance (Pittner 1995; Rouchota 1998; Kärkkäinen 2010;
Kaltenböck 2010); we will return to this issue below.20 Furthermore, placement can
also have a bearing on the function and prosody of theticals: As the work by Ifanti-
dou (1991: 201) and Rouchota (1998: 101–2) suggests, for example, the later a theti-
cal comment clause (e.g., I wonder) or sentence adverbial (aer all) occurs in an
utterance, the more likely it is that it receives a repair or aerthought interpretation.
On thetical grammar
While the placement of theticals is fairly free, it is far from unconstrained,
as has been demonstrated in a number of analyses (e.g., Knowles 1980; Altmann
1981; McCawley 1982; Espinal 1991; Reis 1995; Pittner 1995; Haegeman and Hill
n.d.; and others); their syntax is still largely unclear. Each item or construction has
its own constraints, and some items are restricted essentially to one position with-
in an utterance. In what follows, we will present a few general observations (see
e.g. Espinal 1991: 751–7; Peterson 1999: 238–40; Haegeman and Hill 2009; and
the contributions in Kaltenböck et al. 2010 for more detailed analysis). For some
authors, dierences in positional behavior are taken as a basis for distinguishing
types of theticals, or even for setting theticals o from other kinds of categories.
De Vries (2007: 204–6), for example, proposes distinguishing between appositions
and theticals (proper) mainly on the basis of the fact that the latter have no anchor
and/or no xed position vis-à-vis an anchor unit.
eticals tend to follow their anchor, and Grenoble (2004: 1967) suggests that
this is to be expected: Since they function to particularize or specify an entity, they
can be expected to follow that entity. But this is by no means a requirement. As
we saw in examples presented earlier (e.g., (1d)), theticals may well precede their
anchor, and this also applies to theticals that modify nouns, as we saw in example
(25).
Example (25) also illustrates another feature of theticals. It has been argued in
some form or other that theticals can be placed only at the boundaries of phrasal
constituents of an anchor clause (Emonds 1973: 335; Altmann 1981; Fanselow and
Felix 1987: 31). Accordingly, Jackendo (2002: 256) says that his sentential ad-
verbs, which we consider to be paradigm instances of theticals, “can go in any of
the major breakpoints of the sentence”. Such a claim is in need of revision (see e.g.
Pittner 1995; Tseng 2000; Kaltenböck 2007: 43). While they are most commonly
inserted at the boundaries of phrases, theticals may also cut across phrasal con-
stituents. us, in (25) the thetical is inserted between a demonstrative determiner
and a noun, in (27) between a possessive modier and a noun, and in (28) between
an article and a noun:
(27) Your your <,> uh what’s this called thorax <,> is probably vibrating far too
much. (ICE-GB: s1a-018, 53; Kaltenböck 2007: 41)
(28) Uh in the uhm <,> I think October issue of Computational uh Linguistics
there’s an attempt to do something of this type (ICE-GB: S1a-024, 105;
Kaltenböck 2008: 108)
While many theticals are positionally fairly free, their placement tends to be de-
termined by the discourse functions they serve, and it usually has implications for
the semantic-pragmatic scope relations (Pittner 1995; Kaltenböck 2007; 2010: 248;
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
see below), for the information structure (Taglicht 1984: 22–5; Rouchota 1998;
Dehé and Kavalova 2006: 301), or for the anaphoric reference structure within an
utterance. Consider the following example: Whereas he in (29a) cannot be ana-
phoric to John, in (29b) it may or may not be anaphoric (Knowles 1980: 394–5).
(29) a. John, he believes, will not oer himself as a candidate.
b. John will, he believes, not oer himself as a candidate. (Knowles
1980: 395)
But there are also many constraints on the placement of thetical categories.
e constraints listed below are based on a preliminary survey of English theticals;
more research is urgently required.
(i) Unlike English nominal SG modiers, which precede the head noun with few
exceptions (e.g., relative clauses), appositive theticals (or phrasal appositives) gen-
erally follow their anchor. Accordingly, if the order of the nominal constituents in
(30a) is reversed, as in (30b), then the structure remains the same: It is always the
second NP that is interpreted as a thetical:
(30) a. e whole family — John, Mary and the kids — just disappeared.
b. John, Mary and the kids — the whole family — just disappeared.
(Burton-Roberts 2005, (5))
Unlike the other constraints listed here, (i) appears to be fairly strict. Neverthe-
less, it does not generally apply to theticals that function as nominal modiers.
Modiers such as comment clauses are occasionally found before their nominal
anchor, as in (31a) or in (25) above. Note further that in (31b), her father can be
interpreted as the anchor and a die-hard conservative an appositive thetical (or an
ascriptive supplement in the terminology of Huddleston and Pullum 2002: 1358)
— yet, the anchor follows the thetical:
(31) a. Jim Bob and Fruitbat jokingly suggested that it might be vaguely amusing
to tie in the, if you will, “concept” of the album with a foreign press
conference to promote it. (New Musical Express, 1992, BNC; Brinton
2008: 166)
b. A die-hard conservative, her father refused to even consider the proposal.
(Huddleston and Pullum 2002: 1358)
(ii) Nominal theticals must occur adjacent to their anchor (Emonds 1973: 338;
Grenoble 2004: 1967). In example (32a), of all places species in Paris, hence it
must occur adjacent to in Paris, cf. (32b). But once more, there are a few excep-
tions, especially when there is an indenite anchor unit, as in (32c):
On thetical grammar
(32) a. John wrote a book in Paris, of all places.
b. * In Paris, John wrote, of all places, a book. (Emonds 1973: 338)
c. I met a friend of yours at the party last night — Emma Carlisle.
(Huddleston and Pullum 2002: 1358)
(iii) Unlike restrictive relative clauses, thetical relatives, i.e. non-restrictive (ap-
positive) ones, must immediately follow their anchor (Emonds 1979; Espinal
1991: 753).
(iv) Tag questions may not precede their anchor clause (Knowles 1980: 382–3).
us, (33a) and (33b) are well-formed, but not (33c).
(33) a. John went, didn’t he, to Paris on Sunday?
b. John went to Paris on Sunday, didn’t he?
c. * Didn’t he John went to Paris on Sunday? (Knowles 1980: 382–3)
(v) In the arrangement of pragmatic markers, attention markers such as interjec-
tions always precede orientation markers, i.e. conceptual theticals (see Table 2).
us, Fraser (2009: 893) argues that in examples such as the following, an attention
marker like oh generally precedes an orientation marker like that reminds me:
(34) You regained consciousness about half an hour aer I started talking to you.
Oh, that reminds me, when you came round, you mumbled something about
having heard me calling you. Do you remember that? (Fraser 2009: 893)
To conclude, while the placement of theticals is distinctly more mobile than
that of corresponding SG units, there are a number of constraints, the exact nature
of which is still largely unknown.
. Ellipsis
Our denition of theticals in (7e) of Section 2.1 stipulates that theticals are on the
one hand built on principles of SG but on the other hand they may be elliptic. A
thetical can be self-contained, that is, be entirely in accordance with the rules of
canonical sentence grammar (SG), but it may as well be reduced (or “elliptic”) in a
way that is at variance with requirements of SG.
Also referred to as deletion, omission, reduction, or truncation, ellipsis has
been described in terms of operations such as equi-deletion, gapping, etc. (see
e.g. Shopen 1973). Ellipsis of text pieces whose meaning can be retrieved from the
context can be observed anywhere in language use and language structures. While
being a ubiquitous mechanism, ellipsis is most common in discourse organization
(cf. Rhee 2002): It is a structural characteristic of many thetical constructions (cf.
Biber et al. 1999: 1099). Ellipsis is a complex subject matter, also one that is highly
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
theory-dependent: Whether an information unit is to be analyzed as being ellipti-
cal depends on what kind of non-elliptic structure is postulated by the author or
the model concerned. For the purposes of this discussion, the term is restricted to
cases where a thetical lacks a constituent whose use would be obligatory in a cor-
responding unit of SG.
While there have so far been no detailed empirical investigations into the mat-
ter, it would seem that ellipsis can be of two kinds: It can be due to a gradual
process, whereby frequently used thetical forms are shortened over time, losing
in phonetic and/or morphological substance. Formulae of social exchange in par-
ticular tend to erode over time, thereby acquiring the appearance of elliptic utter-
ances. In this way, an English unit such as Goodbye is a shortened version of God
be with you!, which underwent erosion in the course of its history. Rather than
ellipsis, we will use the term erosion for this kind of gradual process, a term taken
from grammaticalization theory (see Heine and Kuteva 2007, Chapter 2).
Alternatively, ellipsis can be an instantaneous process, in that an SG unit re-
cruited as a thetical is not taken in its full form but rather in some reduced form.
e question, however, is whether ellipsis is in fact an appropriate term to describe
or account for what happens. Compare the examples in (35) with those in (36)
(Kaltenböck 2007: 29–31, 41; Brinton 2008: 10–11). English verbs such as think,
seem, or say require a complement, as (33) illustrates: All three are used as matrix
verbs taking a complement clause (but see 2.7 below; see also Kaltenböck 2009;
2011; Dehé and Wichmann 2010a). In (36), the same verbs occur as theticals with-
out a complement — hence, there can be said to be reduction or ellipsis.
(35) a. I think that the world is at.
b. It seems that the solution is an easy one.
c. He said that Britons could compete and win.
(36) a. e world is at, I think.
b. e solution, it seems, is an easy one.
c. Britons, he said, could compete and win. (ICE-GB; Kaltenböck 2007: 41)
e question of what the structural relationship between (35) and (36) is has
given rise to a range of dierent interpretations, most of them relating to com-
ment clauses such as (36a). A review of all these interpretations would be beyond
the scope of this paper; the reader is referred in particular to Dehé and Kavalova
(2006: 295) and Brinton (2008: 10–14) for discussion.21
As we argue here, the two structures in (35) and (36) are related in that the
matrix clauses of (35) are transferred from SG to thetical grammar (TG), there-
by turning into the theticals of (36) (cf. Section 3.2). is operation can receive
two dierent interpretations. According to the rst, let us call it the retained-
On thetical grammar
complement hypothesis, theticals are no longer subject to restrictions of SG: eir
complement can now be of a dierent nature, such as an independent clause or
sentence. In the wording of Peterson (1999: 233), there is “an anaphoric linkage
between the interpolated clause and its ‘host’, created by ellipsis.”
e second possible interpretation is one in terms of ellipsis, let us call it the
ellipsis hypothesis: In SG, certain arguments of a clause must obligatorily be ex-
pressed; in TG, by contrast, this is not a requirement: Information units whose
content is recoverable from the context can be omitted under appropriate condi-
tions.
Either or both of these consequences can be involved in our example. In the
case of the retained-complement hypothesis, the nature of the complement chang-
es on the way from SG to TG: Being a complement clause of I think in SG (35a), the
world is at is conceptualized and encoded as the main clause in (36a). Alterna-
tively, in the ellipsis hypothesis, the thetical lacks a complement since the meaning
of the latter is recoverable from the context provided by the utterance — in other
words, the “redundant” complement is omitted. An argument in favor of the latter
interpretation is provided by the fact that there are both comment clauses and re-
porting clauses with verbs that do not normally take a complement, e.g. whimper,
smile, or exult, as in the reporting clause of (37). Such examples suggest that the re-
tained-complement hypothesis is not without problems (cf. Biber et al. 1999: 196).
(37) “I’m sorry,” she whimpered. (Biber et al. 1999: 196)
A similar example is provided by tag questions, such as (38). Knowles
(1980: 397) uses the term truncation to refer to the fact that tags lack a verb phrase
constituent. He treats it as a rule that is “phenomenon-specic” in that there do
not seem to be other facts of English which require quite such a rule. Once again,
both accounts sketched above seem possible: Either the proposition Suzanne is
selling her horse is conceptualized as a complement of the tag question, or the tag
question is an utterance that is elliptic since the content of its verb phrase can be
recovered from the situation of discourse.
(38) Suzanne is selling her horse, isn’t she?
Comment clauses, reporting clauses, tag questions, etc. are instances of con-
structional theticals (see Table 1 above). But the same kind of “ellipsis” can also
be observed in many cases of instantaneous theticals. e examples in (39) show
some common types of “ellipsis” in instantaneous theticals:22 Ellipsis concerns the
(copula) verb in (39a), the verb phrase in (39b), and the subject and the copula
verb in (39c). While such kinds of elliptic structures are characteristic of thetical
information units, they are unlikely to be found in SG units.
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
(39) a. e visitors, most of them students, were rather surprised. (Kaltenböck
2007: 30)
b. For those of us who remember nineteen sixty-ve one or two of our
listeners may Tory party uh leadership contests used to be uh as the
cardinals in Rome and leaders would emerge. (ICE-GB: s1b-024, 1; Dehé
2007: 267)
c. e chairman, angry at the delay, demanded a full report. (Kaltenböck
2007: 30)
Ellipsis is a characteristic of many theticals — to the extent that at times it
constitutes the only feature that allows distinguishing theticals from correspond-
ing SG units; at the same time, its exact nature is still largely unclear. For example,
there are cases where ellipsis applies to SG structures but not to theticals. A para-
digm case of (optional) ellipsis in SG is provided by reference identity e.g. between
subjects in coordinated clauses, as in (40a), where the subject of the second clause
is deleted since it agrees with the subject of the rst clause. Yet, an “equi-deletion”
rule does not apply when the second clause is a thetical and-clause (see Table 1),
as in (40b): e subject pronoun he of the thetical clause must not be deleted, even
though it is referentially identical with the subject of the preceding anchor clause23
(Kavalova 2007: 155–6).
(40) a. and an appalling character came and sat down immediately uh uhm in
front of me <,,> and engaged me in conversation.24 (DCPSE: DL-B29,
#758; Kavalova 2007: 155)
b. Emmanuel Shinwell thought and he is aer all a previous Defense
Minister that you can’t have informed opinion on this vital matter without
… (DCPSE: DL-D04, #128; Kavalova 2007: 156)
To conclude, there appears to be a fundamental dierence in the treatment of
ellipsis between SG units and theticals. In more general terms one may say that
theticals are not subject to some syntactic constraints characterizing SG units;
rather, they are shaped primarily by the situation of discourse. But a problem that
needs much further research concerns the question of what exactly accounts for
the structural dierences between SG and TG in the treatment of ellipsis.
. Instantaneous, constructional, and formulaic theticals
ere are a number of typological surveys of English theticals (see especially Es-
pinal 1991: 726–7; Peterson 1999; Dehé and Kavalova 2007b: 1–4; Kaltenböck
2007: 27–31; de Vries 2007: 204; Brinton 2008: 9–10). Other than with the classi-
cation of thetical functions that was the subject of Section 2.3, we are not concerned
with structural typology in this paper. But we need to introduce a distinction that
On thetical grammar
is relevant to the subject matter of the present paper, even if we are not able to deal
with it in as much detail as would be desirable. Consider the examples in (41).
Many of the theticals discussed in this paper, such as the ones in (2) or (3), are
like we can’t hide the fact in (41a). We propose to call them instantaneous theticals:
ey are fully compositional, can be formed freely anytime and anywhere, can be
inserted in most syntactic slots of a sentence, and quite a few of them are uttered
only once and never again. Designed on principles of SG (see 3.2), they exhibit few
constraints in their internal and their external structure.
(41) a. And of course during the nineteenth century we can’t hide the fact Egypt
became very much a hunting ground for agents on behalf of museums in
Europe. (ICE-GB: s2a-052–81)
b. ‘In those days,’ Sue admitted, ‘we were heavily in debt.’ (Huddleston and
Pullum 2002: 1026–7)
c. e therapist’s level tone is bland and neutral — he has, for example,
avoided stressing ‘you’. (Huddleston and Pullum 2002: 1355)
It would however seem that, more oen than corresponding SG units, theti-
cals take the form of “prefabs”, that is, of combinations of words that co-occur
frequently (Erman and Warren 2000), and some of them may undergo what we
loosely refer to as constructionalization (Evans 2007: 370). Constructional theti-
cals, which form our second type, are recurrent patterns or constructions of theti-
cals, being compositional but having some schematic structure and function.
us, the constructional thetical Sue admitted illustrated in (41b) is a reporting
clause that typically consists of a human subject and a speech act verb. A catalog
of common English constructional theticals was presented earlier in Table 1. at
catalog, which is by no means exhaustive, contains a range of syntactically dispa-
rate constructions; but to the extent that they correspond to our denition of (7)
and are each associated with a specic discourse function, they qualify as thetical
categories.
e information unit for example in (39c) illustrates a third type, one that
we propose to call formulaic theticals. ese are non-compositional information
units, that is, their shape is essentially invariable. ey are usually short chunks,
morphosyntactically unanalyzable, tend to be positionally exible and to express
functions that are mostly procedural, and they relate to the situation of discourse
rather than to sentence syntax. Oen these theticals are prosodically integrated
into the anchor utterance as a result of frequent use (e.g. discourse marker uses of
I think, cf. Kaltenböck 2008; Dehé 2009b; Dehé and Wichmann 2010b)
Table 2 provides typical examples of formulaic theticals to be found in Eng-
lish.25 In addition, many of the information units that tend to be classied as dis-
course markers also belong here. Note that some formulaic theticals, in particular
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
primary interjections such as ouch, wow (Ameka 1992), or hesitation markers and
pause llers (Hayashi and Yoon 2006; 2010), do not correspond to property (7e)
of our denition of theticals in that their internal structure does not appear to be
built on principles of sentence grammar.
Table 2. Types of formulaic theticals of English.
Category Examples
Conceptual theticals as it were, for example, if at all, if you will
Formulae of social exchange Good Morning, hello, please
Vocatives Sir!, Waiter!
Imperatives Come on!
Interjections boy, damn, fuck, hell, ouch, pst, um, wow
While both instantaneous and constructional theticals tend to be ignored or
mentioned only in passing in reference grammars, formulaic theticals are usually
distinguished as forming or being part of a separate class of some kind. Classes of
this kind are described as consisting, for example, of minor sentences or sentences
of the minor type (Bloomeld 1962 [1933]: 176; Hockett 1958: 200–1), expressives
(Langacker (2008: 475), or simply of interjections, or interjectives (Doke [1927]
1988: 279–83).
At the same time, there is a fundamental distinction between instantaneous
theticals on the one hand and the other two on the other: Whereas the former are
suggestive of an activity, or a real-time interactional strategy, the last two relate
to discourse grammar as a product, a conventionalized domain consisting of an
inventory of more or less xed information units, stored for reuse in a more or less
frozen or reied form (Du Bois 1985: 362).
. Sentence grammar vs. thetical grammar
e analysis of linguistic discourse is confronted on the one hand with the inten-
tions of interlocutors and discourse purposes; on the other hand it concerns online
speech packaging and, more narrowly, the linguistic tools used for constructing
utterances. We observed in Section 1 that SG and TG form the two main domains
of discourse grammar, and that the two need to be distinguished on account of the
contrasting syntactic, prosodic, and semantic features each is associated with. In
Section 3.1 we will be looking further into this distinction. As we will see in Sec-
tion 3.2, there is also regular interference between the two domains.
On thetical grammar
. Distinguishing properties
We assume that SG is organized in terms of propositional concepts and clauses, and
that the nucleus of the clause is the verb with its argument structure, optionally ex-
tended by peripheral participants (or adjuncts). Its main building blocks are constit-
uent types such as phrases, words, and morphemes plus the syntactic and morpho-
logical machinery to relate these constituents to one another. e notion of SG that
we propose here is a narrow one compared to that of many linguistic models, which
tend to treat theticals more or less as marginal appendages of SG (see Section 4).
ere would in fact be reasons for assigning SG a more central role vis-à-vis
TG. First, it provides a syntactic frame for structuring sentences. Second, it ac-
counts for most of the conceptual information and linguistic material to be found
in texts. ird, it also provides an indispensible anchor for many, even if not for
all categories of TG; for example, it is a sine qua non for the use of tag questions,
appositives, discourse markers, and some other kinds of theticals. And fourth, it
provides the etymological source for theticals, in that the latter are for the most
part taken from units of SG (see 3.2).
Nevertheless, there are reasons not to accord SG any privileged status vis-à-
vis TG, in particular the following. First, like SG units, the information units of
TG typically can form utterances of their own. Second, in spoken discourse, TG
units play an important role in structuring linguistic interaction, being at times
more central to the communicative goals of speakers than SG units. Finally, and
most importantly, each of the two domains has its own principles of organizing
discourse. e major dimension of SG is syntactic hierarchy, which is the central
concern of most mainstream schools of contemporary linguistics (see Section 4).
With its rich repertoire of form-meaning units and its diversied structure of re-
lating these units to one another, SG allows the encoding of virtually all conceptual
information in a coherent and a consistent way.
But TG has one advantage over SG in this respect. It is optimally suited to
attend to the immediate communicative needs of the discourse situation: Rather
than being restricted to the information encoded in the text, it has the entire situ-
ation of discourse in its scope: the speaker, the hearer, and their relation to one
another, to the text, and to the situation in which discourse takes place (see (18)
above). TG thereby adds an extra dimension, described by Espinal (1991) in terms
of a three-dimensional syntactic model, allowing the speaker to temporarily es-
cape the narrow connes of linearity. is is particularly useful for immediate,
spontaneous reactions to the discourse situation; see Section 2.3 for a functional
characterization of this dimension.
e second dierence concerns the constraints on speech packaging. Here,
the benet of TG can be seen in its resilience, where resilience is of two dierent
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
kinds. On the one hand it concerns the external structure of theticals, namely
their mobility in positioning and scope assignment (see 2.4). Although not being
entirely mobile, theticals are generally not restricted to the placement rules of SG,
and this freedom enables the speaker to present information in the most strategic
position of an utterance beyond the rules of SG. On the other hand, this resilience
relates to the internal form of theticals and manifests itself in economy, usually de-
scribed in the relevant literature with reference to ellipsis (see 2.5): By allowing to
omit constituents in a manner and a magnitude that SG does not allow, TG makes
it possible to suppress information that is less central or predictable on account of
the linguistic or extralinguistic discourse situation.
at TG constitutes a separate domain of discourse grammar is also suggested
by the way it is portrayed by the authors who have dealt with it. Dixon (2005: 235–
7), for example, characterizes theticals (parentheticals) in the following way: ey
(a) are most oen short, (b) generally provide simple statements, (c) are very un-
likely to have a negative predicate, or an inherently negative verb (such as forget,
doubt),26 (d) most commonly, their verb is in generic tense, and (e) they exhibit
constraints on which verbs can or cannot occur in thetical comment clauses.
Another feature distinguishing the two domains relates to the ontological na-
ture of information units. As was observed above, SG is organized in terms of
propositional concepts, appropriately enccoded by clauses and sentences. Accord-
ingly, units of SG that do not express or imply full propositions tend to be judged
as incomplete and described in terms of notions such as ellipsis, deletion, trunca-
tion, and the like. TG is also concerned with propositional concepts, especially
since many of the information units built on units of SG take the form of clausal
propositions (see 3.2). But TG appears to be a more exible domain: Its units can
also present other kinds of concepts, such as predicates without arguments, or
with a reduced number of arguments, arguments without predicates, linking con-
cepts without linked material and vice versa, etc. Accordingly, TG utterances may
take a number of dierent forms that units of SG may not, such as that of a phrase
or a single word, of an incomplete main clause, or of a subordinate clause without
main clause (see e.g. the examples in (39)). We treated such utterances in Sec-
tion 2.5 as “elliptic” clauses or sentences since we were restricted to relating them
to corresponding SG structures, being aware that their exact cognitive nature is a
subject of further research.
. Cooptation
We saw in the preceding sections that most theticals have the same form as “cor-
responding” forms of SG. e reason for this is, as we argue in this paper, because
they are the result of a cognitive-communicative operation which we propose to
On thetical grammar
call cooptation. With this term we refer to a packaging strategy whereby a clause,
a phrase, a word, or any other unit is taken from SG and is coopted (or re-dened)
for use as a thetical. Take the following example that we already had in (10). In
(42a), the unit really is an adverb, a constituent of SG. In (42b), the very item re-
ally is not a SG unit; it fullls all the criteria that we proposed for a thetical in (7).
(42) a. Bob is really a poet. (= (10))
b. Bob is a poet, really.
Cooptation is a fully productive operation whereby a unit of SG, such as really
in (42a), is used to serve within the domain of TG: Really is no longer a prosodic or
syntactic part of the clause, and its meaning is also no longer restricted by the rules
of SG but is rather shaped by the situation of discourse, as sketched in Section 2.3.
e exact nature of cooptation as a cognitive operation is still largely unclear and
must be the subject of a separate analysis.27 But on the basis of the evidence avail-
able it would seem that it can be characterized as in (43).
(43) Features characterizing cooptation
a. It is an instantaneous operation whereby a unit of SG is used to serve as
a thetical.28
b. e result is an information unit that is syntactically and prosodically
autonomous, that is, one that corresponds to the denition of theticals
in (7).
c. e meaning of the coopted unit is shaped by its function in discourse.
is may entail e.g. a drastic widening of scope, where widening is not
restricted to the text concerned but relates to the entire situation of
discourse. As Tom Givón (p.c.) suggests, theticals may be structurally
placed into an anchor utterance, but functionally they are in an “anchor
discourse”; their functional scope/range/import is a much wider chunk
of discourse. Accordingly, Traugott and Dasher (2002: 40) conclude
that widening may lead from “scope within proposition” to “scope over
discourse”.
d. Having been coopted for use as a thetical, the unit is freed from
the syntactic and semantic constraints of SG, and it may have the
appearance of an elliptic piece compared to the corresponding structure
of SG (see Section 2.5).
e. But even when coopted as an elliptic piece, the unit can inherit valency
features, although such features may no longer relate to surrounding text
pieces but rather to the situation of discourse in general.
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
. Discussion
In the course of all the research that has been carried out on theticals over the
last decades, a number of important ndings have surfaced. Reviewing all these
ndings would be beyond the scope of this paper; for convenient summaries see
Espinal (1991: 736–41), Grenoble (2004), Blakemore (2006), Dehé and Kavalova
(2007b), Kaltenböck (2007), and Brinton (2008: 7–14). Our interest here is with
the question of how the framework outlined in this paper diers from others that
have been proposed and what advantages this framework oers in understanding
the domain of thetical grammar.
ere are some positions that are hard to reconcile with the position main-
tained here. is applies most of all to a line of research where it is argued in some
way or other that TG can be reduced to principles of SG (e.g., Jackendo 1972;
Ross 1973; Emonds 1973; 1976; Lako 1974; McCawley 1982; Hand 1993; 2002;
Corver and iersch 2002; Potts 2002; Schelout et al. 2004; Tucker 2005; Axel
and Kiziak 2006; Kiziak 2007).
To be sure, theticals are almost invariably built on principles of SG and in-
teract with units of SG via cooptation (see 3.2). But we are not aware of any ap-
proach based exclusively on conventional syntax of any kind that would be able to
oer a satisfactory account for the properties of theticals discussed in Section 2.
eoretical notions that have been invoked to account for the syntactic behavior
of theticals are in particular movement, sliing, extraction, extraposition, and the
like (see Shaer and Frey 2004 for a dicussion on le-detached theticals). “Move-
ment” in a broader sense has in fact gured in some way or other in many ap-
proaches of dierent theoretical orientations. Terms such as sliing (“sentence li-
ing”, Ross 1973), extraction (Postal 1998; Schelout et al. 2004; Axel and Kiziak
2006; Kiziak 2007), and extraposition (cf. Reis 1995; Crystal 1997; Couper-Kuhlen
and ompson 2009) have been proposed to deal with operations whereby an ele-
ment is moved out of its canonical position to some other place within a sentence
or utterance. But rather than movement, cooptation appears to be a cognitive op-
eration whereby existing material is exploited for new functions, and rather than
taking place within one and the same domain, it involves two dierent domains of
discourse organization.
Obviously, the framework proposed here is most closely related to what, fol-
lowing Haegeman (1988; 1991), tends to be referred to summarily as “the radical
orphanage approach”. is approach assumes that theticals fall outside the domain
of SG, or clausal syntax for that matter (Haegeman 1988; 1991; Espinal 1991: 740;
Peterson 1999; Burton-Roberts 1999; 2005; Shaer 2003; Shaer and Frey 2004; Aver-
intseva-Klisch 2008), or “should be kept separate from the tree structure, related
to their anchors by some dierent notational device” of syntactic representation
On thetical grammar
(Huddleston and Pullum 2002: 1353–4). ere are a number of dierences among
these authors, especially on the question of where exactly theticals are to be located.
For most of these authors, theticals are not constituents of any structure but are in-
tegrated into the anchor utterance in some kind of “post-syntactic procedure”, e.g.,
via a discourse-governed process of linearization (Dehé and Kavalova 2006) or at
the pragmatic level of utterance interpretation (Espinal 1991; Haegeman 1991; Bur-
ton-Roberts 1999; Averintseva-Klisch 2008: 236), as is perhaps particularly obvious
in examples of stand-alone theticals, such as the insubordinated clause in (44).
(44) If you don’t mind [said by a speaker moving past neighbors to take his seat
in a crowded cinema]. (Haegeman 1988; Espinal 1991: 740)
A problem that is at the core of many syntactic studies concerns the conict
between linearization and hierarchical integration: eticals are linearly ordered
but do not correspond to rules of syntactic hierarchy in their external structure
(e.g., Haegeman 1988; 1991; Espinal 1991; Burton-Roberts 2005; Dehé and Ka-
valova 2007b; Steinbach 2007; Kiziak 2007; Kaltenböck 2007; see also Fabb 1990).
is problem does not exist in approaches which assume that SG and TG are built
on dierent principles but interact in jointly producing linearly ordered texts.
In order to solve this problem, Espinal (1991: 741–57) proposes a three-di-
mensional space where the phrase marker of the anchor sentence (i.e., the SG unit)
and that of the disjunct sentence (the thetical) are assigned to dierent dimensions
of structural representation. While these phrase markers share no syntactic node,
they are related to one another by means of “precedence structural relations”, that
is, via linearization: e structural axis connecting the two dimensions forms the
linear intersection between the dimensions, and the relation between the two
phrase markers is not one of syntactic dependence but rather takes place at the
conceptual level of representation or at the level of discourse structure. In the case
of theticals that take the form of comment clauses, such as the one we had in (36a)
(e world is at, I think), Espinal (1991: 748) asserts that the linguistic meaning of
the anchor (the world is at) “is projected into the empty argument position of the
verb of the parenthetical” (think) in the nal process of utterance interpretation.
Among the dierent perspectives that have been discussed there is on the one
hand one distinguishing between two contrasting kinds of narrative style: Whether
the distinction between SG and TG can be related to that between a reportive and
an expressive narrative style (Baneld 1973) is an issue that would be worth being
looked into but cannot be decided here. On the other hand there is a perspective
according to which there is a fundamental dierence between two types of theti-
cals, namely integrated and non-integrated ones (Reis 1995; Shaer and Frey 2004;
Steinbach 2007: 53–84; Dehé and Kavalova 2007b: 6–7; Kavalova 2007: 157–63).
is issue has yielded an enormous body of research and insights into the syntac-
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
tic behavior of theticals. While it is still largely unclear how this distinction can be
explained,29 it would seem that one of the main forces underlying the distinction
is constructionalization (see 2.6).
As these few remarks suggest, most works on the analysis of theticals have
focused on their external syntax. We diverge from some earlier authors especially
in maintaining that any attempt at reducing TG to principles of SG syntax, or rel-
egating them to some “post-syntactic” level of analysis is unlikely to do justice to
the external and/or the internal structure of theticals. Rather, we assume that this
structure is shaped by the network of components characterizing the situation of
discourse (see Section 2.3, (18)). Such components include most of all the source
of information, attitudes of speakers, speaker-hearer interaction, the discourse
setting, and world knowledge (cf. Grenoble 2004; Blakemore 2006; Kaltenböck
2007; and others).
Where we dier from the authors mentioned in the preceding paragraphs in
particular is, rst, that we argue that TG cannot be reduced to being a structural
appendage of SG but rather constitutes a domain of linguistic behavior in its own
right. Second, we observed in Section 2.6 that theticals are not all of the same
kind; rather, there are at least three main types, namely instantaneous, construc-
tional, and formulaic theticals. Whereas instantaneous theticals are suggestive of
an activity, the last two relate to discourse grammar as a conventionalized domain
consisting of an inventory of more or less xed information units. And nally, we
argue that rather than syntactic operations such as movement, sliing, extraction,
and extraposition, the mechanism leading to the creation of theticals is coopta-
tion, whereby the grammatical resources available in one domain, namely SG, are
recruited for designing information units in another domain, i.e. that of TG.
. Conclusions
In the present paper, a wide range of works were discussed. at theticals cannot
be reduced to canonical principles of sentence grammar has been pointed out by a
number of authors since the late 1980s. e ndings made in this paper dier most
from those of previous works as follows: First, we argue that thetical grammar forms
a distinct domain that, while interacting in a number of ways with the domain of
sentence grammar, cannot be accounted for by means of some “notational device”,
or in terms of some kind of post-syntactic or pragmatic appendage, of “processing
constraints”, or as a “production phenomenon” (cf. Espinal 1991; Peterson 1999: 240;
Burton-Robert 1999). Rather than being determined by the morphosyntactic struc-
ture of the sentence, the meaning of theticals is shaped by a network of conceptual
components that we referred to in Section 2.3 as the situation of discourse.
On thetical grammar
Second, while there is no shortage of typological surveys of thetical categories,
the distinction that we introduced in Section 2.6 may be of help in understanding
why a number of the generalizations that were proposed earlier do not extend to
a larger range of thetical phenomena. In quite a few studies, comment clauses and
non-restrictive relative clauses were taken as a basis for understanding the nature
of theticals. But these are both instances of constructional theticals (see Table 1),
while the contribution of instantaneous and formulaic theticals to discourse or-
ganization is still largely terra incognita. Especially the role played by formulaic
theticals such as the ones listed in Table 2 is in need of much further research.
ird, we were concerned with the interference between sentence grammar
and thetical grammar. is interference was described in earlier works in terms
of notions such as movement, sliing, extraction, and extraposition. A review of
the various controversies that have taken place in connection with this issue sug-
gests, however, that none of these notions is really satisfactory, and we argued in
Section 3.2 that the cognitive-communicative mechanism leading to the creation
of new theticals is cooptation, whereby the grammatical resources available in the
domain of sentence grammar are recruited for designing information units in the
domain of thetical grammar. e cooptation hypothesis is able to account for why
quite a number of linguistic expressions (e.g., between you and me, frankly, really,
in other words, well, etc.) occur simultaneously in both domains and are each as-
sociated with two contrasting sets of properties.
A number of the positions reviewed in Section 4 were shaped by a preoccu-
pation with specic subsets of theticals.30 At the same time we are aware that the
discussion in the present paper also shows a number of limitations. It is biased in
particular in favor of some kinds and uses of theticals, while other thetical cat-
egories, such as formulae of social exchange, vocatives, interjections, including
hesitation markers, llers, etc., could be mentioned only in passing and must be
the subject of a separate study.
eticals are elements which the speaker (or writer) presents as separate from
sentence grammar in order to signal what Dik (1997: 396) called a “higher level
orientation function2. ey allow the speaker to “step out” of the connes of the
linearity of communication to some extent by creating a kind of second plane
communication, not unlike “asides” on stage. is plane can be inserted spontane-
ously virtually anywhere and therefore lends itself particularly well to situation-
specic, metacommunicative information. But the plane needs to be signaled to
the listener as such prosodically by separate tone units, pauses, etc., and by the
suspension or loosening of constructional constraints and ensuing mobility.
Creating theticals is thus a fairly elementary human activity which can be ac-
tivated at all times and in any context; and yet, we still know little about what
motivates and constrains this activity. Accordingly, the generalizations proposed
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
so far have not been enough to understand the nature and the entire potential of
theticals, nor do they allow us to predict when and why they are employed. As we
argue in this paper, linguistic discourse operates in at least two dierent domains,
namely that of sentence grammar and thetical grammar. ese two domains —
even though distinct — nevertheless also interact in a number of ways. Perhaps
the main form of interaction can be seen in cooptation, whereby the resources
available in SG are exploited for use in TG.
ree issues in particular were central to the discussion in this present paper.
e rst concerns the ontological status of theticals. Having generally been treated
as a peripheral phenomenon, we argue that they play a signicant role in dis-
course organization, and some forms of linguistic interaction are framed primarily
in terms of formulaic thetical expressions. at theticals form fairly monolithic
blocks is suggested by the fact that, rst, they cannot be split (see 2.4), second, they
are usually separated from the rest of an utterance both syntactically and prosodi-
cally and, third, they can form utterances of their own.
e second issue relates to the place of theticals within a theory of grammar or
discourse. As we observed in Section 4, this issue has been discussed extensively,
and controversially, centering around the question of whether theticals can be rec-
onciled with established notions of syntax, as is argued for most of all by adherents
of integrated-syntax approaches, whether they cannot be reduced to syntax and
hence have to be dealt with in some kind of post-syntactic procedure, as claimed
in some form or other in approaches of the radical orphanage type (see Section 4),
or whether they constitute a domain of their own, as argued for in this paper.
eticals are, in the wording of Dehé and Kavalova (2007b: 4), a “Janus-faced lot”,
on the one hand exhibiting various kinds of interference with sentence grammar,
on the other hand constituting a structurally independent domain.
e third issue concerns the taxonomic status of theticals: Do really all phe-
nomena that we discussed in this paper form one unied domain? We have an-
swered this question in the armative on the basis of the denition proposed in
(7) of Section 2.1. At the same time we also observed there that some information
units were included in this category even if they do not conform in every aspect
to the denition: Syntactic and prosodic boundaries do not necessarily coincide
(see e.g. Dehé 2007). Should such units be excluded from the domain, or is there
need for a broader denition and, if yes, what should such a denition look like
without being largely vacuous? As we mentioned in Section 2.6, but were not able
to discuss in detail, the more frequently a thetical is used, the more it tends to lose
its syntactic and prosodic autonomy. e result is that a number of theticals are
elusive to analysis in terms of discrete categorization.
On thetical grammar
Acknowledgements
We wish to express our gratitude to a number of colleagues who have been of help in writing
this paper, providing many critical comments, in particular to Felix Ameka, Jim Bennett, Lau-
rel Brinton, Ulrike Claudi, Bob Dixon, Wolfgang Dressler, Jack DuBois, Martin Haspelmath,
Jack Hawkins, Christa König, Haiping Long, Maarten Mous, Gábor Nagy, Heiko Narrog, Fritz
Newmeyer, Klaus-Uwe Panther, John Joseph Perry, Seongha Rhee, Kyung-An Song, Danjie Su,
Shanti Ulfsbjorninn, Arie Verhagen, as well as to the participants of the International Confer-
ence on Final Particles, held in Rouen on 27–28 May 2010, to the participants of the conference
Beyond Dichotomies, held in Budapest on 25–26 October, 2010, to the participants of the con-
ference on Competing Motivations, held in Leipzig on 23–25 November 2010, as well as to our
colleagues at Leiden University for their critical comments. Furthermore, we wish to thank Tom
Givón and two anonymous reviewers for many valuable comments on an earlier version of this
paper. Finally, the rst-named author is also grateful to the Korean Ministry of Education, Sci-
ence and Technology for generously having sponsored the research leading to this paper within
its World Class University Program (Grant Nr. R33–10011). e third co-author is greatly in-
debted to SOAS, University of London and the Alexander von Humboldt Foundation for their
generous support.
Abbreviations
ACC = accusative, oblique case-marking sux; BNC = British National Corpus; C = concord
(noun class marker); CP = thetical; DCPSE = Diachronic Corpus of Present-day Spoken Eng-
lish; DO = discourse organization; F = feminine; ICE-GB = International Corpus of English,
British component; OE = Old English; OED = Oxford English Dictionary; PA = plural addressee
marker; POSS = possessive; Q = question marker; SG = sentence grammar, singular; TG = theti-
cal grammar; VOC = vocative case marker; 1, 2, 3 = rst, second, third person.
Notes
. In the text examples presented in this paper, thetical units are printed in bold.
. is does not exhaust the list of dierences. Another dierence consists in the fact that (1a)
exhibits a productive use of between whereas (1b) is suggestive of a largely formulaic expression.
. Instead of “anchor”, the terms host or frame are used by a number of authors. Our preference
for the term anchor is simply because “host” has been applied to a number of other purposes,
e.g. in grammaticalization theory, where “host” refers to the stable or invariable part of a com-
bination, or in morphology, where the term applies to the word to which a clitic is attached (see
Huddleston and Pullum 2002: 1351). Cf. also the term “anchored” parentheticals as used by
Wichmann (2001: 179).
. For particularly complex, multiply encapsulated theticals, see the examples of syntactic amal-
gams in Lako (1974).
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
. ompson (2002: 143) also uses the term “parenthetical” but in a more restricted sense, e.g.
for frequent complement-taking predicate phrases that become epistemic adverbial phrases,
such as I guess, I think.
. e term thetical must not be confused with that of “thetic” statement (Kuroda 1972; Sasse
1987; 2006; Lambrecht 1994); both terms are derived from the Ancient Greek verb tithēmi “to
put, pose” (Sasse 2006). etic statements contrast with categorical ones especially in having
neither an intonational nor an informational dichotomy, they are all-focus utterances that assert
a state of aairs with no topic-comment or foreground-background structure. Unlike thetical
categories, thetic statements or judgments are sentence grammar units, which include e.g. im-
personal and existential statements such as It’s raining or God exists.
. In view of the syntactic independence of some of the categories, Biber et al. (1999: 1082)
propose a type of nonclausal unit that they call inserts for English units such as sorry, oh, thanks,
or well.
. Note that ompson (2002: 143) restricts the term parenthetical to frequently used “epis-
temic adverbial phrases” occurring in medial or nal position of their anchor clause.
. “Discourse grammar” is composed of all the linguistic means that are available for construct-
ing spoken or written texts (see 3.1). An analysis of it would be beyond the scope of the present
paper and will be the subject of a separate study.
. See Evans (2007), Heine et al. (2011) for a treatment of insubordination.
. A problem notoriously associated with the study of theticals is that many authors have a
specic subset, such as comment clauses in mind when dening or proposing generalizations
about them.
. In doing so, we follow the suggestion of an anonymous reviewer of this paper.
. In the preposition instead of, by contrast, instead is not a thetical, not being prosodically set
o from the following information unit.
. With regard to pauses, an anonymous reviewer of this paper rightly asks “what if someone
simply pauses because they forget a word?” We are concerned here not with unique acts but
rather with recurrent patterns of language use.
. ere are cases which require further analysis, in that they are syntactically compatible with
both a sentence grammar and a thetical analysis, e.g., I think John is a nk (see Kaltenböck 2009,
2011; Dehé and Wichmann 2010a: 38). In speech, such cases can presumably be disambiguated
via prosodic analysis.
. An anonymous reviewer of this paper argues that this last observation suggests that one
should not speak of “thetical” as a discrete category but, rather, of a cline of theticality.
. e “comment” function is taken by a number of authors to be part of the denition of theti-
cals. For Rouchota (1998: 105), for example, theticals (parentheticals) are “syntactically uninte-
grated elements which are separated from the anchor clause by comma intonation and function
as comments”.
On thetical grammar
. Concerning the distinction between instantaneous and constructional theticals, see Sec-
tion 2.6.
. It goes without saying that grammaticalized, e.g. inectional evidential markers, as they can
be found in many languages, are not theticals.
. We are not concerned here with syntactic scope in the sense of c-command relations. Se-
mantic-pragmatic scope concerns the “topic” to which the predication expressed by the thetical
applies (see Kaltenböck 2010: 248).
. As pointed out to us by Laurel Brinton (p.c.) there is the following problematic case, which
can be interpreted both as SG structure (via ellipsis of the complementizer that) and as a pre-
posed thetical: I think the world is at. See our discussion of this issue in Section 2.2; see also
Dehé and Wichmann 2010a.
. Concerning the distinction between instantaneous theticals and grammaticalized construc-
tions, see 2.6. As we observed above, the notion “ellipsis” is not independent of the theory that
one wishes to adopt; accordingly, not every reader may agree with the classication of types of
ellipsis proposed below.
. For good reasons, Kavalova (2007: 155) therefore observes that this may be indicative of the
syntactic independence of the thetical from its anchor.
. e symbol “<,,>” stands for an extended pause.
. An analysis of the categories listed in Table 2 will be the subject of a separate study.
. As Dixon (2005: 236) adds, however, “double negatives” are possible: Mary is, I don’t doubt,
a good doctor, Tomorrow is, don’t forget, a public holiday. But this does not conclude the list of
possible negative theticals, e.g., Jane is, don’t you think, no longer in love with him.
. We are grateful to an anonymous reviewer of this paper for having drawn our attention to
problems associated with cooptation. Among the questions that we are not able to answer at
the present stage of research it is most of all the following: Are theticals conceptualized and de-
signed within SG and subsequently coopted in TG, or do speakers conceive them in the domain
of TG and subsequently draw on appropriate expressions in SG?
. e reason for not referring to this process as “change” is because cooptation is in principle
a unique, instantaneous operation. It is only when this operation is performed frequently that it
may lead to grammatical change (see Section 2.6).
. Kaltenböck (2007: 33) observes that non-clausal theticals are more prone to being syntacti-
cally integrated into the anchor than clausal ones. at it is most of all relatively short theticals
such as comment clauses, reporting verbs, and tag questions that may be prosodically integrated
into some adjacent constituent has in fact been established in a number of empirical studies (see
Crystal 1969; Taglicht 1998; Wichmann 2001; Peters 2006; Dehé 2007; 2009b).
. For example, for Huddleston and Pullum (2002: 895), theticals are “expressions which can
be appended parenthetically to an anchor clause but which also have a non-parenthetical use in
which they take a declarative content clause as complement.”
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
References
Aarts, B. & C. F. Meyer (eds.). 1995. e Verb in Contemporary English. Cambridge: Cambridge
University Press.
Aijmer, Karin. 1972. Some Aspects of Psychological Predicates in English. Stockholm: Almqvist
and Wiksell.
Aijmer, Karin 1997. “I think” — an English modal particle. In T. Swan and O. Jansen Westvik
(eds.), Modality in Germanic Languages: Historical and Comparative Perspectives, 1–47.
Berlin, New York: Mouton de Gruyter.
Altmann, Hans. 1981. “Formen der Herausstellung” im Deutschen: Linksversetzung, Rechtsverset-
zung, Freies ema und verwandte Konstruktionen (Linguistische Arbeiten 106). Tübingen:
Niemeyer.
Ameka, Felix. 1992. Interjections: e universal yet neglected part of speech. Journal of Prag-
matics 18. 101–18.
Arroyo, José Luis Blas. 2011. From politeness to discourse marking: e process of gragmatical-
ization of muy bien in vernacular Spanish. Journal of Pragmatics 43. 855–74.
Astruc, L. 2005. e Intonation of Extra-Sentential Elements in Catalan and English. Cambridge:
University of Cambridge dissertation.
Averintseva-Klisch, Maria. 2008. To the right of the clause: Right dislocation vs. aerthought. In
Fabricius-Hansen, Cathrine and Wiebke Ramm 2008, ‘Subordination’ versus ‘Coordination’
in Sentence and Text: A Cross-Linguistic Perspective (Studies in Language Companion Series
98), 217–39. Amsterdam, Philadelphia: Benjamins.
Axel, Katrin & Tanja Kiziak. 2006. Contributing to the extraction/parenthesis debate: Judge-
ment studies and historical data. Paper presented at the International Conference on Lin-
guistic Evidence, 2–4 February 2006, University of Tübingen.
Baneld, Ann. 1973. Narrative style and the grammar of direct and indirect speech. Foundations
of Language 10. 1–39.
Barth-Weingarten, Dagmar & Elizabeth Couper-Kuhlen. 2002. On the development of nal
though: A case of grammaticalization? In Wischer, Ilse and Gabriele Diewald (eds.), New
Reections on Grammaticalization, 345–361. Amsterdam, Philadelphia: Benjamins.
Bayer, Klaus. 1973. Verteilung und Funktion der sogenannten Parenthese in Texten. Deutsche
Sprache 1. 64–115.
Biber, Douglas, S. Johansson, G. Leech, S. Conrad & E. Finegan. 1999. Longman Grammar of
Spoken and Written English. London: Longman.
Blakemore, Diane. 1987. Semantic Constraints on Relevance. Oxford: Blackwell.
Blakemore, Diane. 2005. And-parentheticals. Journal of Pragmatics 37. 1165–81.
Blakemore, Diane. 2006. Divisions of labour: e analysis of parentheticals. Lingua 116. 1670–
87.
Bloomeld, Leonard. 1962 [1933]. Language. London: George Allen & Unwin.
Bolinger, Dwight. 1989. Intonation and its Uses: Melody in Grammar and Discourse. London:
Edward Arnold.
Boye, Kasper & Elisabeth Engberg-Pedersen (eds.). 2010. Usage and Structure: A Festschri for
Peter Harder. Berlin: Mouton de Gruyter.
Boye, Kasper & Peter Harder 2007. Complement-taking predicates: Usage and linguistic struc-
ture. Studies in Language 31(3). 569–606.
On thetical grammar
Brandt, Margareta. 1996. Subordination und Parenthese als Mittel der Informationsstrukturier-
ung in Texten. In Motsch, Wolfgang (ed.), Ebenen der Textstruktur: Sprachliche und kom-
munikative Prinzipien, 211–40. Tübingen: Niemeyer.
Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse Func-
tions (Topics in English Linguistics 19). Berlin and New York: Mouton de Gruyter.
Brinton, Laurel J. 2008. e Comment Clause in English: Syntactic Origins and Pragmatic Devel-
opment (Studies in English Language). Cambridge: Cambridge University Press.
Burton-Roberts, Noel. 1999. Language, linear precedence and parentheticals. In Collins, Peter &
David Lee, e Clause in English, 33–52. Amsterdam, Philadelphia: Benjamins.
Burton-Roberts, Noel. 2005. Parentheticals. Encyclopedia of Language and Linguistics, 2nd edn,
vol. 9, 179–82. Amsterdam: Elsevier.
Bybee, Joan L. 2007. Frequency of Use and the Organization of Language. Oxford: Oxford Uni-
versity Press.
Corrigan, Roberta, Edith A. Moravcsik, Hamid Ouali & Kathleen M. Wheatley. 2009. Formulaic
Language. Volume 1: Distribution and Historical Change (Typological Studies in Language
82). Amsterdam, Philadelphia: Benjamins.
Corum, Claudia. 1975. A pragmatic analysis of parenthetic adjuncts. Chicago Linguistic Society
11. 133–41.
Corver, N. & C. iersch. 2002. Remarks on parentheticals. In van Oostendorp, M. & E. An-
agnostopoulou (eds.), Progress in Grammar: Articles at the 20th Anniversary of the Com-
parison of Grammatical Models Group in Tilburg. Online pubication, http://www.meertens.
knaw.nl/books/progressingrammar/
Couper-Kuhlen, Elizabeth & Sandra A. ompson. 2009. On assessing situations and events in
conversation: ‘extraposition’ and its relatives. Discourse Studies 10(4). 443–67.
Crystal, David. 1969. Prosodic Systems and Intonation in English. Cambridge: Cambridge Uni-
versity Press.
Crystal, David. 1997. A Dictionary of Linguistics and Phonetics, 4th edn. Cambridge, MA: Black-
well.
D’Avis, F. J. 2005. Über Parenthesen. In d’Avis, F. J. (ed.), Deutsche Syntax: Empirie und eorie,
259–79.Gothenburg: Acta Universitatis Gothoburgensis.
de Vries, Mark. n.d. e syntax of nonsubordination: Parenthesis, appositions and gras. http://
odur.let.rug.nl/~dvries/pdf/projectvoorstel-website.pdf
de Vries, Mark. 2007. Invisible constituents? Parentheses as B-merged adverbial phrases. In
Dehé, Nicole & Yordanka Kavalova, Parentheticals (Linguistics Today 106), 203–34. Am-
sterdam, Philadelphia: Benjamins.
Dehé, Nicole. 2007. e relation between syntactic and prosodic parenthesis. In Dehé, Nicole
& Yordanka Kavalova, Parentheticals (Linguistics Today 106), 261–85. Amsterdam, Phila-
delphia: Benjamins.
Dehé, Nicole. 2009a. Parentheticals. In Cummings, Louise (ed.), e Pragmatics Encyclopedia,
307–8. London, New York: Routledge.
Dehé, Nicole. 2009b. Clausal parentheticals, intonational phrasing, and prosodic theory. Journal
of Linguistics 45(3). 569–615.
Dehé, Nicole & Yordanka Kavalova. 2006. e syntax, pragmatics, and prosody of parenthetical
what. English Language and Linguistics 10. 289–320.
Dehé, Nicole & Yordanka Kavalova. 2007a. Parentheticals (Linguistics Today 106). Amsterdam,
Philadelphia: Benjamins.
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
Dehé, Nicole & Yordanka Kavalova. 2007b. Introduction. In Dehé, Nicole & Yordanka Kavalova,
Parentheticals (Linguistics Today 106), 1–22. Amsterdam, Philadelphia: Benjamins.
Dehé, Nicole & Anne Wichmann. 2010a. Sentence-initial I think (that) and I believe (that): Pro-
sodic evidence for use as main clause, comment clause and discourse marker. Studies in
Language 34(1). 36–74.
Dehé, Nicole & Anne Wichmann. 2010b. e multifunctionality of epistemic parentheticals
in discourse: Prosodic cues to the semantic-pragmatic boundary. Functions of Language
17(1). 1–28.
Diewald, Gabriele. 2011. Pragmaticalization (dened) as grammaticalization of discourse func-
tions. Linguistics 49(2). 365–90.
Dik, Simon C. 1997. e eory of Functional Grammar, Part 2: Complex and Derived Construc-
tions (Functional Grammar Series 21). Berlin, New York: Mouton de Gruyter.
Dixon, R.M.W. 1979. Ergativity. Language 55. 59–138.
Dixon, R.M.W. 2005. A Semantic Approach to English Grammar. Oxford: Oxford University
Press.
Doke, Clement M. [1927] 1988. Textbook of Zulu Grammar, 6th edn. Cape Town: Maskew Mill-
er Longman.
Du Bois, John. 1985. Competing motivations. In Haiman, John (ed.). Iconicity in Syntax (Typo-
logical Studies in Language 6), 343–66.Amsterdam: Benjamins.
Emonds, Joseph. 1973. Parenthetical clauses. In Corum, C. et al. You take the high node and I’ll
take the low node, 333–47.Chicago: Linguistic Society.
Emonds, Joseph. 1976. A Transformational Approach to English Syntax. New York: Academic
Press.
Emonds, Joseph. 1979. Appositive relatives have no properties. Linguistic Inquiry 10(2). 211–43.
Erman, Britt & Beatrice Warren. 2000. e idiom principle and the open choice principle. Text
20. 29–62.
Espinal, M. 1991. e representation of disjunct constituents. Language 67. 726–62.
Evans, Nicolas 2007. Insubordination and its uses. In Nicolaeva, Irina (ed.), Finiteness: eoreti-
cal and Empirical Foundations, 366–431. Oxford: Oxford University Press.
Fabb, N. 1990. e dierence between English restrictive and non-restrictive relative clauses.
Journal of Linguistics 26. 57–78.
Fanselow, Gisbert & Sascha Felix. 1987. Sprachtheorie: Volume 2: Die Rektions- und Bindung-
stheorie (Studien zur deutschen Grammatik 33). Tübingen: Narr.
Fischer, Kerstin (ed.). 2006. Approaches to Discourse Particles. Amsterdam: Elsevier.
Fortmann, Christian. 2006. e complement of verba dicendi parentheticals. In Miriam Butt
& Tracy Holloway King (eds.), Proceedings of the LFG06 Conference, Universität Konstanz.
Konstanz: CSLI Publications
Fortmann, Christian. 2007. e complement of reduced parentheticals. In Dehé, Nicole & Yor-
danka Kavalova, Parentheticals (Linguistics Today 106), 89–119. Amsterdam, Philadelphia:
Benjamins.
Frank-Job, Barbara. 2006. A dynamic-interactional approach to discourse markers. In Fischer,
Kerstin (ed.) 2006. Approaches to Discourse Particles. Amsterdam: Elsevier. 359–74.
Fraser, Bruce. 2007. e English contrastive discourse marker instead. In Butler, Christopher
S., Raquel Hidalgo Downing & Julia Lavid (eds.), Functional Perspectives on Grammar and
Discourse: In Honour of Angela Downing (Studies in Language Companion Series 85), 301–
12. Amsterdam, Philadelphia: Benjamins.
Fraser, Bruce 2009. Topic orientation markers. Journal of Pragmatics 41. 892–8.
On thetical grammar
Giacalone Ramat, Anna & Caterina Mauri 2009. Introduction to the workshop “Grammaticaliza-
tion between semantics and pragmatics”. Typescript, Pavia, February 2009.
Givón, T. 2009. e Genesis of Syntactic Complexity: Diachrony, Ontogeny, Neuro-Cognition, Evo-
lution. Amsterdam, Philadelphia: Benjamins.
Grenoble, Lenore. 2004. Parentheticals in Russian. Journal of Pragmatics 36(11). 1953–74.
Grewendorf, G. 1988. Aspekte der deutschen Syntax. Tübingen: Narr.
Grosz, Barbara J. & Candace L. Sidner. 1986. Attention, intentions, and the structure of dis-
course. 12(3). 175–204.
Grosz, Barbara and Yael Ziv. n.d. Centering, global focus, and right dislocation. Typescript.
Günthner, Susanne. 1993. ‘… weil — man kann es ja wissenschalich untersuchen‘: diskursprag-
matische Aspekte der Wortstellung in WEIL-Sätzen. Linguistische Berichte 143. 37–59.
Günthner, Susanne. 1999. Entwickelt sich der Konzessivkonnektor obwohl zum Diskursmarker?
Grammatikalisierungstendenzen im gesprochenen Deutsch. Linguistische Berichte 180.
409–46.
Günthner, Susanne. 2000. From concessive connector to discourse marker: e use of obwohl
in everyday German interaction. In Elizabeth Couper-Kuhlen & Bernd Kortmann (eds.),
Cause-Condition-Concession-Contrast: Cognitive and Discourse Perspectives, 439–68. Ber-
lin: Mouton de Gruyter.
Günthner, Susanne & K. Mutz. 2004. Grammaticalization vs. pragmaticalization? e develop-
ment of pragmatic markers in German and Italian. In W. Bisang, N. Himmelmann and B.
Wiemer (eds.), What Makes Grammaticalization? A Look from its Fringes and its Compo-
nents. Berlin: Mouton de Gruyter.
Haegeman, Liliane. 1991. Parenthetical adverbials: the radical orphanage approach. In Chiba,
S. et al. (eds.), Aspects of Modern English Linguistics: Papers Presentedto Masatomo Ukaji on
his 60th Birthday. Tokyo: Kaitakusha. 232–54.
Haegeman, Liliane & Virginia Hill. n.d. e syntactization of discourse. Typescript. http://www.
socsci.ulster.ac.uk/comms/onli2010/paper/Haegeman&Hill%20Syntactization%20of%20
discourse..pdf
Hand, Michael. 1993. Parataxis and parentheticals. Linguistics and Philosophy 16. 495–507.
Hayashi, Makoto & Kyung-Eun Yoon. 2006. A cross-linguistic exploration of demonstratives in
interaction: With particular reference to the context of word-formulation trouble. Studies
in Language 30. 485–540.
Hayashi, Makoto & Kyung-Eun Yoon. 2010. A cross-linguistic exploration of demonstratives in
interaction: With particular reference to the context of word-formulation trouble. In Ami-
ridze, Nino, Boyd H. Davis, and Margaret Maclagan (eds.), Fillers, Pauses and Placeholders
(Typological Studies in Language 93), 33–65. Amsterdam, Philadelphia: Benjamins.
Heine, Bernd, Gunther Kaltenböck & Tania Kuteva. 2011. Accounting for insubordinated claus-
es. Typescript.
Heine, Bernd & Tania Kuteva. 2002. World Lexicon of Grammaticalization. Cambridge: Cam-
bridge University Press.
Heine, Bernd & Tania Kuteva. 2007. e Genesis of Grammar: a Reconstruction (Studies in the
Evolution of Language 9). Oxford: Oxford University Press.
Hockett, Charles F. 1958. A Course in Modern Linguistics. New York: MacMillan.
Homann, L. 1998. Parenthesen. Linguistische Berichte 175. 299–328.
Howe, Chad. 2008. Parentheticals. A review of Dehé and Kavalova 2007. LINGUIST List
19.1761, June 2, 2008.
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
Huddleston, R. & G. K. Pullum. 2002. e Cambridge Grammar of the English Language. Cam-
bridge: Cambridge University Press.
Ifantidou, Elly. 2001. Evidentials as Relevance (Pragmatics and Beyond IV:6). Amsterdam, Phila-
delphia: Benjamins.
Ifantidou-Trouki, Elly. 1993. Sentential adverbs and relevance. Lingua 90(1–2). 69–90.
Izutsu, Mitsuko Narita & Katsunobu Izutsu. 2010. Hanging or back-shiing? : e rise of the
nal particle but and their comparables in Japanese. Paper presented at the International
Conference on Final Particles, 27–28 May, 2010, Rouen, France.
Jackendo, Ray. 1972. Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT
Press.
Jackendo, Ray. 2002. Foundations of language: Brain, meaning, grammar, evolution. Oxford:
Oxford University Press.
Jucker, Andreas H. & Yael Ziv. (eds.) 1998. Discourse Markers: Description and eory. Amster-
dam, Philadelphia: Benjamins.
Kaltenböck, Gunther. 2005. Charting the boundaries of syntax: A taxonomy of spoken paren-
thetical clauses. Vienna Working Papers 14(1). 21–53. http://www.univie.ac.at/Anglistik/
Views0501ALL.pdf.
Kaltenböck, Gunther. 2007. Spoken parenthetical clauses in English. In Dehé, Nicole and Yor-
danka Kavalova, Parentheticals (Linguistics Today, 106), 25–52. Amsterdam, Philadelphia:
Benjamins.
Kaltenböck, Gunther. 2008. Prosody and function of English comment clauses. Folia Linguistica
42(1). 83–134.
Kaltenböck, Gunther. 2009. Initial I think: Main or comment clause? Discourse and Interaction
2(1). 49–70. http://homepage.univie.ac.at/gunther.kaltenboeck/
Kaltenböck, Gunther. 2010. Pragmatic functions of parenthetical I think. In Kaltenböck, Gun-
ther, Gudrun Mihatsch & Stefan Schneider (eds.), New Approaches to Hedging, 243–72.
Emerald Publishers.
Kaltenböck, Gunther. 2011. Linguistic structure and use: Explaining diverging evidence. e
case of clause-initial I think. In Schönefeld, Doris (ed.), Converging evidence: methodologi-
cal and theoretical issues for linguistic research. Amsterdam: Benjamins.
Kaltenböck, Gunther, Gudrun Mihatsch & Stefan Schneider (eds.). 2010. New Approaches to
Hedging. Emerald Publishers.
Kärkkäinen, Elise. 2003. Epistemic Stance in English Conversation: A Description of its Interac-
tional Functions, with a Focus on I think. Amsterdam, Philadelphia: Benjamins.
Kärkkäinen, Elise. 2010. Position and scope of epistemic phrases in planned and unplanned
American English. In Kaltenböck, Gunther, Gudrun Mihatsch and Stefan Schneider (eds.),
New Approaches to Hedging, 208–41. Emerald Publishers.
Karlsson, F. (ed.). Papers from the ird Scandinavian Conference of Linguistics. Turku: Textlin-
guistics Research Group, Academy of Finland.
Kavalova, Yordanka. 2007. And-parenthetical clauses. In Dehé, Nicole & Yordanka Kavalova.
Parentheticals. (Linguistics Today 106), 145–72. Amsterdam, Philadelphia: Benjamins.
Kiziak, Tanja. 2007. Long extraction or parenthetical insertion? Evidence from judgement stud-
ies. In Dehé, Nicole & Yordanka Kavalova, Parentheticals. (Linguistics Today 106), 121–44.
Amsterdam, Philadelphia: Benjamins.
Knowles, John. 1980. e tag as a parenthetical. Studies in Language 4(3). 379–409.
Kruisinga, E. 1932. A Handbook of Present-Day English. Part II. Groningen: Noordho.
On thetical grammar
Kuroda, S.-Y. 1972. e categorical and the thetic judgment: Evidence from Japanese syntax.
Foundations of Language 9. 153–85.
Kutik, Elanah J., Cooper, William E. & Boyce, Suzanne. 1983. Declination of fundamental fre-
quency in speakers’ production of parenthetical and main clauses. Journal of the Acoustical
Society of America 73(5). 1731–1738.
Lako, George. 1974. Syntactic amalgams. Chicago Linguistic Society 10. 321–44.
Lambrecht, Knud. 1994. Information Structure and Sentence Form. Cambridge: Cambridge Uni-
versity Press.
Lambrecht, Knud. 1996. On the formal and functional relationship between topics and voca-
tives: Evidence from French. In Goldberg, Adele E. (ed.), Conceptual Structure, Discourse
and Language (CSLI Pulications), 267–288. Stanford: Center for the Study of Language and
Information.
Langacker, Ronald W. 2008. Cognitive Grammar: A Basic Introduction. Oxford: Oxford Univer-
sity Press.
Lewis, Diana M. 2011. A discourse-constructional approach to the emergence of discourse
markers in English. Linguistics 49(2). 415–43.
McCawley, James D. 1982. Parentheticals and discontinuous constituent structure. Linguistic
Inquiry 13. 91–106.
McCawley, James D. 1988. e Syntactic Phenomena of English (Two volumes). Chicago: Uni-
versity of Chicago Press.
McGloin, Naomi H. &Yumiko Konishi. 2010. From connective particle to sentence-nal par-
ticle: A usage-based analysis of shi ‘and’ in Japanese. Language Sciences 32. 563–78.
Mazeland, Harrie. 2007. Parenthetical sequences. Journal of Pragmatics 39. 1816–69.
Michaelis, Laura A. & Knud Lambrecht. 1996. e exclamative sentence type in English. In
Goldberg, Adele E. (ed.), Conceptual Structure, Discourse and Language (CSLI Publica-
tions), 375–89. Stanford: Center for the Study of Language and Information.
Mindt, I. 2003. Is I think a discourse marker? In Mengel, E. et al., Proceedings Anglistentag 2002,
Bayreuth, 473–83. Trier: WVT.
Morgan, Jerry. 1973. Sentence fragments and the notion ‚sentence‘. In Kachru, Braj, Robert Lees,
Yakov Malkiel, Angelina Pietrangeli & Sol Saporta 1973. Issues in Linguistics, 719–51. Ur-
bana: University of Illionois Press.
Mulder, Jean & Sandra A. ompson. 2008. e grammaticization of but as a nal particle in
English conversation. In Laury, Ritva (ed.), Crosslinguistic Studies of Clause Combining: e
Multifunctionality of Conjunctions, 179–204. Amsterdam: John Benjamins.
Mulder, Jean, Sandra A. ompson & Cara Penry Williams. 2009. Final but in Australian Eng-
lish conversation. In Collins, Peter & Pam Peters (eds.), Comparative Grammatical Studies
in Australian and New Zealand English. Amsterdam: John Benjamins.
Nakayama, Toshihide & Kumiko Ichihashi-Nakayama. 1997. Japanese kedo: Discourse genre
and grammaticization. In Sohn, Ho-min & John Haig (eds.), Japanese/Korean Linguistics,
vol. 6, 607–618. Stanford: CSLI Publications.
Nespor, Marina & Irene Vogel. 1986. Prosodic Phonology. (Studies in Generative Grammar 28).
Dordrecht: Foris.
Newmeyer, Frederick J. 2010. What conversational English tells us about the nature of grammar:
A critique of ompson’s analysis of object complements. In Boye, Kasper & Elisabeth
Engberg-Pedersen (eds.), Usage and Structure: A Festschri for Peter Harder, 3–43. Berlin:
Mouton de Gruyter.
Nosek, J. 1973. Parenthesis in modern colloquial English. Prague Studies in English 15. 99–116.
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
Ohori, Toshio. 1995. Remarks on suspended clauses: a contribution to Japanese phraseology. In
Shibatani, Masayoshi & Sandra A. ompson (eds.), Essays in Semantics and Pragmatics,
201–218. Amsterdam: John Benjamins.
Peters, J. 2006. Syntactic and prosodic parenthesis. Proceedings of the International Conference
on Speech Prosody, Dresden, 2–5 May 2006.
Peterson, Peter. 1999. On the boundaries of syntax: Non-syntagmatic relations. In Peter Collins
& David Lee (eds.), e Clause in English: In Honour of Rodney Huddleston, 229–250. Am-
sterdam, Philadelphia: John Benjamins.
Petola, Niilo. 1983. Comment clauses in present-day English. In Kajanto, Iiro et al. (eds.), Studies
in Classical and Modern Philology, 101–13. Helsinki: Suomalainen, Tiedeakatemia.
Pinker, Stephen & Michael T. Ullman. 2002. e past-tense debate: e past and future of the
past tense. Trends in Cognitive Sciences 6. 456–63.
Pittner, Karin. 1995. Zur Syntax von Parenthesen. Linguistische Berichte 156. 85–108.
Polinsky, Maria. 1996. Situation perspective: On the relations of thematic roles, discourse cat-
egories, and grammatical relations to gure and ground. In Goldberg, Adele E. (ed.), Con-
ceptual Structure, Discourse and Language (CSLI Publications), 401–19. Stanford: Center
for the Study of Language and Information.
Postal, P. 1998. ree Investigations of Extraction. Cambridge, MA: MIT Press.
Potts, Christopher. 2002. e syntax and semantics of as-parentheticals. Natural Language and
Linguistic eory 20. 623–89.
Potts, Christopher. 2003. Conventional implicatures: A distinguished class of meanings. In G.
Ramchand and C. Reiss (eds.), e Oxford Handbook of Linguistic Interfaces. Oxford: Ox-
ford University Press.
Prevost, Sophie. 2011. A propos from verbal complement to discourse marker: A case of gram-
maticalization? Linguistics 49(2). 391–413.
Reinhart, Tanya. 1983. Point of view in language — the use of parentheticals. In Gisa Rauh (ed.),
Essays on Deixis, 169–194. Tübingen: Narr.
Rhee, Seongha. 2002. From silence to grammar: Grammaticalization and ellipsis in Korean.
Paper presented at the conference on New Reections on Grammaticalization II, University
of Amsterdam, April 3–6, 2002.
Rhee, Seongha. 2004. From discourse to grammar: Grammaticalization and lexicalization of
rhetorical questions in Korean. In Gordon Fulton, William J. Sullivan & Arle R. Lommel
(eds.), LACUS: Forum XXX: Language, ought and Reality, 413–23. Houston: Lacus.
Reis, Marga. 1995. Extractions from Verb-Second clauses in German? In Lutz, U. & J. Pafel, J.,
On Extraction and Extraposition in German. Benjamins, Amsterdam.
Ross, John Robert. 1973. Sliing. In Maurice Gross, Morris Halle & Marcel-Paul Schützenberger
(eds.), e Formal Analysis of Natural Languages: Proceedings of the First International Con-
ference, 133–69. e Hague, Paris: Mouton.
Rouchota, Villy. 1998. Procedural meaning and parenthetical discourse markers. In Jucker, An-
dreas H. & Yael Ziv 1998. Discourse Markers: Descriptions and eory, 97–126. Amsterdam,
Philadelphia: Benjamins.
Sacks, Harvey, Emanuel A. Scheglo & Gail Jeerson. 1974. A simplest systematics for the orga-
nization of turn-taking for conversation. Language 50. 696–735.
Sar, K. 1986. Relative clauses in a theory of binding and levels. Linguistic Inquiry 17(4). 663–89.
Sasse, Hans-Jürgen. 1987. e thetic/categorical distinction revisited. Linguistics 25. 511–580.
Sasse, Hans-Jürgen. 2006. eticity. In G. Bernini & M. L. Schwartz, Pragmatic Organization of
Discourse in the Languages of Europe, 255–308. Berlin and New York: Mouton de Gruyter.
On thetical grammar
Scheglo, Emanuel A. 1998. Reections on studying prosody in talk-in-interaction. Language
and Speech 41. 235–63.
Schelout, Carla, Peter-Arno Coppen & Nelleke Oostdijk. 2004. Finite comment clauses in
Dutch: A corpus-based approach. Journal of Linguistics 16. 331–49.
Schirin, Deborah. 1987. Discourse Markers (Studies in Interactional Sociolinguistics 5). Cam-
bridge: Cambridge University Press.
Schneider, Stefan. 2007a. Reduced parenthetical clauses in Romance languages. In Dehé, Nicole
& Yordanka Kavalova, Parentheticals (Linguistics Today, 106), 237–58.Amsterdam, Phila-
delphia: Benjamins.
Schneider, Stefan. 2007b. Reduced Parenthetical Clauses as Mitigators: A Corpus Study of Spoken
French, Italian and Spanish. Amsterdam, Philadelphia: Benjamins.
Schourup, Lawrence. 1999. Discourse markers. Lingua 107. 227–65.
Shaer, Benjamin. 2003. An ‘orphan’ analysis of long and short adjunct movement in English.
In Garding, G. & M. Tsujimira (eds.), WCCFL 22 Proceedings, 450–63. Sommerville, MA:
Cascadilla Press.
Shaer, Benjamin & Werner Frey. 2004. ‘Integrated’ and ‘non-integrated’ le-peripheral elements
in German and English. ZAS Papers in Linguistics 35(2). 465–502.
Shopen, Timothy. 1973. Ellipsis as grammatical indeterminacy. Foundations of Language 10.
65–77.
Stainton, Robert. 1995. Non-sentential assertion and semantic ellipsis. Linguistics and Philoso-
phy 18. 281–96.
Stainton, Robert. 2004. In defence of non-sentential assertion. In Zoltan Szabo 2005. Semantics
vs. Pragmatics, 383–457. Oxford: Oxford University Press.
Stenström, A.-B. 1995. Some remarks on comment clauses. In Aarts, B. & C. F. Meyer (eds.), e
Verb in Contemporary English, 290–9. Cambridge: Cambridge University Press.
Steinbach, Markus. 2007. Integrated parentheticals and assertional complements. In Dehé,
Nicole & Yordanka Kavalova, Parentheticals (Linguistics Today 106), 53–87. Amsterdam,
Philadelphia: Benjamins.
Stowell, Tim. 2005. Appositive and parenthetical relative clauses. In Hans Broekhuis, Norbert
Corver, Jan Koster, Riny Huybregts & Ursula Kleinhenz, Organizing Grammar: Linguistic
Studies in Honor of Henk van Riemsdijk, 608–17. Berlin, New York: Mouton de Gruyter.
Stubbs, Michael. 1983. Discourse Analysis: e Sociolinguistic Analysis of Natural Language and
Culture. Oxford: Blackwell.
Svensson, J. 1976. Report indicators and other parentheticals. In Karlsson, F. (ed.), Papers from
the ird Scandinavian Conference on Linguistics, 369–80. Turku: Textlinguistics Research
Group, Academy of Finland.
Taglicht, J. 1984. Message and Emphasis: On Focus and Scope in English. London: Longmans.
Taglicht, J. 1998. Constraints on intonational phrasing in English. Journal of Linguistics 34.
181–211.
ompson, Sandra A. 2002. ‘Object complements’ and conversation: Towards a realistic ac-
count. Studies in Language 26(1). 125–64.
ompson, Sandra A. & Anthony Mulac. 1991a. A quantitative perspective on the grammati-
cization of epistemic parentheticals in English. In Traugott, Elizabeth C. & Bernd Heine
(eds.), Approaches to grammaticalization, vol. 1 (Typological Studies in Language 19:1),
313–29. Amsterdam, Philadelphia: Benjamins.
ompson, Sandra A. & Anthony Mulac. 1991b. e discourse conditions for the use of the
complementizer that in conversational English. Journal of Pragmatics 13. 237–251.
Gunther Kaltenböck, Bernd Heine and Tania Kuteva
Traugott, Elizabeth C. 1995. e role of discourse markers in a theory of grammaticalization.
Paper presented at the 12th International Conference on Historical Linguistics, Manches-
ter, England, August 1995.
Traugott, Elizabeth C. 2003. From subjectication to intersubjectication. In Hickey, Raymond
(ed.), Motives for Language Change, 124–39. Cambridge: Cambridge University Press.
Traugott, Elizabeth C. & Richard B. Dasher. 2002. Regularity in semantic change. Cambridge
Studies in Linguistics 96). Cambridge: Cambridge University Press.
Traugott, Elizabeth C. & Bernd Heine (eds.). 1991a. Approaches to grammaticalization, vol. 1
(Typological Studies in Language 19:1). Amsterdam, Philadelphia: Benjamins.
Traugott, Elizabeth C. & Bernd Heine (eds.). 1991b. Approaches to grammaticalization, vol. 2
(Typological Studies in Language 19:2). Amsterdam, Philadelphia: Benjamins.
Tseng, Jesse. 2000. Linearizing parentheticals in German: e limitations of phrase structure.
ESSLLI Workshop on Linguistic eory and Grammar Implementation, Tübingen, August
17, 2000.
Tucker, Gordon. 2005. Extending the lexicogrammar: Towards a more comprehensive account
of extraclausal, partially clausal and non-clausal expressions in spoken discourse. Language
Sciences 27(6). 679–709.
Urmson, J. O. 1952. Parenthetical verbs. Mind 61. 480–96.
van Bogaert, Julie. 2011. I think and other complement-taking mental predicates: A case of and
for constructional grammaticalization. Linguistics 49(2). 295–332.
Vandenbergen, Anne-Marie Simon & Dominique Willems. 2011. Crosslinguistic data as evi-
dence in the grammaticalization debate: e case of discourse markers. Linguistics 49(2).
333–64.
Waltereit, Richard. 2002. Imperatives, interruption in conversation and the rise of discourse
markers: A study of Italian guarda. Linguistics 40. 987–1010.
Waltereit, Richard. 2006. e rise of discourse markers in Italian: A specic type of language
change. In Fischer, Kerstin 2006, Approaches to Discourse Markers. (Studies in Pragmatics
1), 61–76. Amsterdam: Elsevier.
Watts, Richard 2003. Politeness. Cambridge: Cambridge University Press.
Wichmann, Anne. 2001. Spoken parentheticals. In Aijmer, K. (ed.), A Wealth of English: Studies
in Honour of Goran Kjellmer, 171–93.Gothenburg: Gothenburg University Press.
Wichmann, Anne. 2004. e intonation of Please-requests: A corpus-based study. Journal of
Pragmatics 36(9). 1521–49.
Wilson, Deidre & Dan Sperber. 1993. Linguistic form and relevance. Lingua 90(1–2). 1–25.
Wischer, Ilse. 2000. Grammaticalization versus lexicalization: ‘Methinks’ there is some confu-
sion. In Fischer, Olga, Anette Rosenbach & Dieter Stein (eds.), Pathways of Change: Gram-
maticalization in English, 355–70. Amsterdam, Philadelphia: John Benjamins.
Wischer, Ilse & Gabriele Diewald (eds.). 2002. New Reections on Grammaticalization. Amster-
dam, Philadelphia: Benjamins.
On thetical grammar
Authors addresses
Gunther Kaltenböck
Department of English
University of Vienna
Spitalgasse 2–4
1090 Vienna, Austria
gunther.kaltenboeck@univie.ac.at
Bernd Heine
Institute for African Studies
University of Cologne
50923 Cologne, Germany
Tania Kuteva
Heinrich-Heine-Universität Düsseldorf
Anglistik III — Geb 23.21
Universitätsstrasse 1
40225 Düsseldorf, Germany
