Summary
#e article examines how the function and the distribution of pragmatic markers in spoken 
discourse are reflected by prosodic means. #e results of this preliminary analysis show that the 
intonation largely depends on the discourse function of a marker, on the one hand, and the 
speaker’s perception of its importance for the ongoing discourse. A vague division can be made 
between the prosodic patterns of those discourse markers which are strongly involved in the 
organisation and monitoring of the discourse, and those whose function is more pragmatic.
 
Key words: linguistics, discourse analysis, intonation
 Povzetek
V članku analiziram, kako se funkcija in položaj pragmatičnih označevalcev v govoru odražata 
v stavčni intonaciji. Rezultati preliminarne analize kažejo, da je stavčna intonacija močno 
odvisna od diskurzne funkcije označevalca na eni strani in govorčeve presoje, kako pomemben je 
označevalec za potek diskurza. V grobem lahko ločimo med intonacijskim vzorcem, ki je značilen 
za tiste diskurzne označevalce, ki imajo bolj organizacijsko in opazovalno vlogo v diskurzu ter 
tistimi, ki imajo bolj pragmatično vlogo.
 
Ključne beside: jezikoslovje, besediloslovje, intonacija
 DOI: 10.4312/elope.4.1-2.43-55

In every spoken interaction speakers use a number of different lexical items
1
 which at first 
sight do not contribute much to the contents of conversation, but are indispensable for the 
organization, structuring and monitoring of discourse, a smooth turn-taking system, as well 
as the expression of speakers’ attitude to the message. Such lexical items are in English well, 
now, oh, but, I mean, you know, as a matter of fact, to mention only a few. In the literature on 
discourse and conversation analysis these lexical items are often referred to by different names, 
including discourse particles, discourse markers, interactional signals, discourse connectives, pragmatic 
markers, or sentence connectives. Some of them are used more often with reference to written 
language, others to spoken. #e diversity of names arises from the fact that these lexical items 
morphologically belong to different parts of speech, are not obligatory and perform different 
discourse and pragmatic functions. 
What is then the purpose of analysing discourse markers? Firstly, discourse markers perform an 
important role in establishing cohesion in speech. #ey often have the anaphoric and cataphoric 
character thus pointing backward and forward in discourse at the same time. Secondly, they are used 
for organizing and monitoring the topic development. As such they open or close a topic, mark 
topic boundaries and attract listeners’ attention. #ey introduce reformulated utterances and make 
reference to shared knowledge. Finally, on a more pragmatic level, they may function as markers of 
the speakers’ view points, their emotional reactions, or to tone down the utterance’s effect.
#e main question in analysing discourse markers is how speakers distinguish between different 
discourse and pragmatic functions of the same lexical item. In other words, how speakers know 
when well is used as a topic opening or closing device or as an introduction to a reformulated 
utterance, or when I mean is a face saving expression or a marker that the following statement is 
a rewording or a clarification of the previous one. 
One answer to the question is the position of the marker in the utterance and  discourse. But 
what clues are available to listeners when the marker occupies the same position in the discourse 
but performs different pragmatic functions? I believe that in such cases intonation often becomes 
significant.  
For this preliminary analysis of the relation between pragmatic markers and their prosodic treatment 
I chose a sample of the BBC TV series As Time Goes By. #e choice of the material was governed by 
the fact that the script contains numerous pragmatic markers and that the actors are aware of their 
discourse function and know how to pronounce them in order to express that function. 
 Lexical item is a cover term used with reference to single words, phrases or even clauses.

Schiffrin, in her seminal work on discourse markers gives an operational definition of them 
as “sequentially dependent elements which bracket units of talk” (1987, 31) and a theoretical 
definition of them as “members of a functional class of verbal (and non-verbal) devices which 
provide contextual coordinates for ongoing talk” (ibid., 41). #e sequential dependence means that 
markers are items which function on a discourse level and are not dependent on smaller linguistic 
units. 
#e list of discourse markers is relatively open and contains lexical items which belong to different 
parts of speech: and, but, or are conjunctions; so can be a conjunction or an adverb; now and 
then are adverbs, whereas oh and well can be adverb, interjection, filler or particle. T o complicate 
matters even more, the discourse markers also include phrases: in general, or even clauses: you 
know. #us it seems reasonable to treat discourse markers according to the functions they perform 
in a discourse and not according to their morphological or syntactical characteristics. 
In Stenström’s model of spoken interaction
2
, markers play important roles as topic organizers and 
monitors, as well as turn-taking devices. #us she makes a distinction between discourse markers 
and interactional signals. Both refer to the same set of lexical items, but the difference lies in their 
function and position in the discourse. “Interactional signals are used to start, carry on and terminate 
the conversation” (ibid., 61), whereas “discourse markers are used to organize and hold the turn 
and to mark boundaries in the discourse” (ibid., 63). #e best way to distinguish between them is 
to view them as “filling a gap in the exchange” (ibid., 60). As such they can make up a turn of their 
own or they may fill a slot in a turn. Stenström underlines the fact that both discourse markers and 
interactional signals can fill more than one gap in the exchange and more than one slot in the turn, 
or can do different things in different places, as well as different things in the same place (ibid., 61).
Interactional signals play an important role in a smooth exchange of turns. As such they can 
appeal for and provide feedback, respond or involve the listener in conversation. Discourse 
markers, on the other hand, serve to start or end a conversation, introduce a topic, subtopic, 
digression or the resumption of the old topic.
 
With the view of encompassing both discourse and pragmatic functions of markers, Carter 
and McCarthy introduce the term pragmatic markers to refer to a “class of items which operate 
outside the structural limits of the clause and which encode speakers’ intentions and interpersonal 
meanings” (2006, 208). Pragmatic markers thus include discourse markers, stance markers, 
hedges and interjections. 
Carter and McCarthy explain discourse markers as words, phrases and clauses which “link segments 
of the discourse to one another in ways which reflect choices of monitoring, organisation and 
management exercised by the speaker” (ibid.,). Table 1 presents the most typical words, phrases 
 Stenström’s model of spoken interaction consists of five levels: act, turn, move, exchange and transaction, of which act 
is the lowest level and transaction the highest.

and even clauses which function in the organisation and monitoring of discourse in English.
Discourse Markers
Organizing Monitoring
Open&-
close
Sequence
Topic
Boundary
Focus Reformulate
Shared
Knowledge
Responding
so;
(all) right;
then;
now;
good;
well;
OK;
anyway;
fine;
lovely;
great.
and;
and then;
finally;
first (of all);
firstly;
for a start;
going back to;
in general;
in the end;
in the first 
place;
last of all;
lastly;
next;
on top of that;
second;
secondly;
so;
there again;
third(ly);
to sum up;
what’s more
OK;
so;
yeah;
and;
right;
anyway.
hey;
listen;
look;
oh;
well;
anyway;
so.
as I was saying;
as it were;
I mean;
if you like;
in a manner of 
speaking;
in other words;
not to say;
or rather;
so to speak;
strictly 
speaking;
that’s to say;
to put it in 
another way;
to put it 
bluntly/mildly;
well.
you see;
see;
you know.
right;
all right;
I see;
good;
great;
fine;
OK.
 Stance markers, hedges and interjections are also pragmatic markers whose functions are to express 
the speaker’s attitude towards the message (i.e. stance markers), to tone down the utterance 
in order not to sound too blunt (i.e. hedges) and to express different emotional reactions to 
the utterance or the situation (i.e. interjections). Table 2 presents lexical items which are most 
typically used as stance markers, hedges and interjections in English.

Stance Markers Hedges Interjections
actually;
admittedly;
amazingly;
basically;
certainly;
clearly;
confidentially;
doubtless;
essentially;
frankly;
to be frank;
fortunately;
honestly;
to be honest;
hopefully;
ideally;
if you ask me;
I’m afraid;
I must admit;
I must say;
I think;
in fact;
indeed;
literally;
naturally;
no doubt;
obviously;
of course;
predictably;
putting (to put) it mildly/bluntly;
(quite) rightly;
really;
sadly;
seriously;
(I’m) sorry;
strictly speaking;
surprisingly;
thankfully;
to tell you the truth;
understandably;
undoubtedly;
unfortunately.
apparently;
arguably:
by any chance;
I think;
just (about) kind of;
like;
maybe;
perhaps;
presumably;
probably;
roughly;
sort of;
surely.
bother;
crikey;
damn;
God;
goodness (me);
gosh;
(good) heavens;
hooray;
jeez;
ooh;
oh no;
oops;
ouch;
ow;
ugh;
tut-tut;
whoops;
wow;
yippee;
yuk.


Intonation performs grammatical, discourse and pragmatic functions in speech. #e first mainly 
concerns the division of speech into intonation phrases (IP) and the location of the nuclear 
syllable; the second and the third are concerned with different pitch movements whose discourse 
function is to express the anaphoric and cataphoric references, as well as contrast in meaning, 
while their pragmatic function is to express speakers’ attitudes towards the message and their 
emotional reactions to the message or situation.
#e fact that the same lexical items are used to express both discourse and  pragmatic functions 
is an adequate reason for a closer prosodic analysis with a view of finding an interface between 
their distribution in discourse, on the one hand, and their discourse and pragmatic roles, on the 
other.
 Carter and McCarthy (2006, 539) claim that prosodic information helps to distinguish 
between discourse markers and other parts of speech or clauses. According to them discourse 
markers often occupy their own intonation phrase and are accompanied by brief pauses. 
Although this may sometimes be the case, it is nonetheless too unreliable to be regarded as 
the main distinction between the discourse and non discourse marking of lexical items. 
If we take, for example, the expression I mean in examples (1) and (2), it is clear that in (1) 
it is part of the clause structure and as such non discourse marking, while in (2) it has the 
parenthetical function within a sentence and is thus regarded as a discourse marker:
(1) I mean what I say.
(2) I mean, wouldn’t it be better to postpone the meeting.
But when it comes to the division of the two examples into intonation phrases, speakers always 
have a choice between the marked or unmarked version. It is thus reasonable to expect that the 
unmarked version of example (1) would consist of one intonation phrase, as in (1a), while in the 
marked version the speaker may insert the intonation phrase division after I mean, as in (1b), and 
thus emphasise his intention:
(1a)  I mean what I say.
4
(1b)  I mean | what I say.
 Tonality is the system of dividing speech into intonation phrases.
 The underlined syllables are nuclear or tonic syllables.

Similarly, in example (2), the speaker may separate the discourse marker I mean from the rest of the 
clause in order to emphasise the rewording or the previous utterance or to clarify it, as in (2a), or 
decide for only one intonation phrase in which I mean is not given any special emphasis, as in (2b):
(2a)  I mean | wouldn’t it be better to postpone the meeting.
(2b)  I mean, wouldn’t it be better to postpone the meeting.
#e relationship between tonality and discourse markers is not always straightforward; the 
chances for a discourse marker to be separated from the rest of the clause by an intonation phrase 
mostly depend on the position of the marker in the discourse (i.e. initial, medial or final in an 
utterance or a turn), as well as its discourse function (i.e. organising or monitoring discourse). But 
even there speakers have a choice to express their ideas in such a way as to guarantee successful 
communication even if it means overriding general guidelines.
 When discourse markers are treated as separate intonation phrases, they contain the nuclear 
syllable where a particular pitch movement or tone is realised. #e pitch height of the tone can 
also be significant. #e decision with which tone and at which pitch height a discourse marker 
will be realised also depends on the position and function of the marker and will be discussed in 
the following.
 #e purpose of this is to discuss different functions of discourse markers and their prosodic 
realisations. #e discourse markers are divided into those which are mainly involved in the 
organization of discourse in terms of opening and closing topics or conversation, expressing 
relationship of sequence between parts of discourse and focusing attention, diverting, shifting and 
resuming a topic, as well as those which enable speakers to monitor and manage the discourse.
 Discourse markers play an important role in the organisation of discourse. #ey are particularly 
frequent in opening and closing a conversation and in opening, re-opening, closing or temporarily 
closing a topic. In all these functions most of the discourse markers will indeed have their own 
intonation phrase since they represent a transition from one topic to another or mark the 
beginning or the end of a conversation. #ey are attention seeking devices and as such deserve a 
prosodic treatment of their own.
According to Brazil’s intonation model (1997) the fall (() is used to introduce information 
which is unknown to the addressee, while the high key (#) expresses contrast to the message in 
the previous intonation phrase or general knowledge.

In example (3) there are two discourse markers which function as discourse closer (well) and 
opener (now). In both cases the speaker treats them as separate intonation phrases and in both 
cases uses falling tones uttered in high key:
(3) Jean: #(Well, | that’s all finished. || #( Now, | what have you been chatting about?
Similarly in examples (4) and (5) the discourse markers so and anyway are used as topic and 
subtopic boundaries, respectively. In (4) Margaret uses so to pick up on information she has 
learnt and which was interrupted by ordering drinks, while in (5) Lionel uses anyway to mark 
the boundary between his behaviour to his ex-wife and the fact that she has suggested to meet 
him. #ey both use separate intonation phrases for the markers and pronounce them in high key 
and with the falling tones:
(4)  Lionel: You didn’t like champagne.
 Margaret: I didn’t like lots of things. | #( So, | you’re married again. || How long 
 ago did that happen?
(5)  Jean: Have you shown an interest in her?
 Lionel: Good God, | no! #( Anyway, | she says she’s in London for a few days 
 and might we meet for a drink.
Another distribution of discourse markers where there is a very high possibility for them to 
be treated as separate intonation phrases is when they are used to focus the attention of the 
listener on something that is of high importance. In example (6) Jean uses the marker look to 
draw Lionel’s attention to an important fact, namely that she cannot get dressed by 8:30. She 
pronounces it as an intonation phrase of its own, in high key and with the falling tone:
 
(6)  Jean: #( Look, | you know what I mean. || I can’t possibly be ready for 8:30.
 #e process of monitoring the discourse mainly concerns reformulations and monitoring shared 
knowledge. #e function of reformulations is to change the wording in such a way that it 
better fits the context of interaction or speakers’ intentions. Whether a discourse marker used 
in this function will be treated as a separate intonation phrase is less predictable than in the case 
of discourse openers or closers. #e choice of tone is also less predictable. Our analysis shows a 
pattern in which the discourse markers well and I mean are most frequently used in this function 
and are as such treated as separate intonation phrases with either falling or level intonation, as in 
examples (7) and (8):

(7)  Sandy: #ey’re never her own, are they?
 Judy: No way! || ( Well, | at least I hope not.
(8)  Jean: Oh, credit me with some taste! || I" mean | he’s very, very good looking 
 but if he has a brain he did a very good job of hiding it.
     
In monitoring shared knowledge speakers check the state of common knowledge for which 
they use expressions you know, you see and see. According to Carter and McCarthy (2006, 539) the 
difference between you know, on the one hand, and you see or see, on the other, is that the former 
assumes that the listener may not have the same knowledge as the speaker, while the latter presumes 
that the speaker and the listener share the same knowledge or have the same point of view. 
#e distribution of these monitors is usually at the end or in the middle of a turn and they are 
often treated as separate intonation phrases. As to the pitch movement, they can be said with 
either rising or falling tones; if the speaker wants to elicit a response from the listener, the rising 
tone is more appropriate, as in example (9):
(9)  Judy: … I’m trying, as discretely as I can, to push Alistair into Sandy’s 
 direction. I mean, she’s a smashing girl. Once you get past his ego, he’s not 
 a bad chap. You & see?
 Jean: I suppose I do.
If the speaker does not want a response from the listener, the falling tone is usually used, as in 
example (10):
(10)  Lionel: Dull nights on television, | you (know. 
Although Carter and McCarthy (2006) do not treat question tags as discourse markers, 
Stenström (1994) does mention them as lexical items which may function as discourse markers 
or interactional signals. And indeed they do perform a monitoring function. Depending on the 
intonation, they may be used to engage the listener in the conversation and invite agreement or 
disagreement with the speaker. In example (11) the rising intonation on the tag clearly invites 
the listener to express an opinion:
(11)  Sandy: #ey’re never her own, | & are they?
 Judy: No way! Well, at least I hope not.
#e tag in example (12), on the other hand, is not felt to be an invitation to respond, but if the 
listener does say something it is usually in agreement with the speaker’s opinion:

(12)  Lionel: My God! She can knock it back, | (can’t she?
   Jean: She must’ve had two bottles to herself.
 Hedges, stance markers and interjections are those lexical items which also stand out of 
the syntactic structure of a clause and which do not contribute much to the main topic of 
conversation; in other words, their absence does not change the meaning or the grammatical 
acceptability of the utterance. But they are nonetheless indispensable expressions if speakers want 
to modify their wording. 
 Spontaneous speech is an online process, which means that listeners hear everything that speakers 
intend to say even if it is not perfectly formulated. Sometimes they may, halfway through the 
utterance, realise that the wording may sound too blunt or assertive so they decide to tone it 
down by using a hedge. 
When it comes to the division into intonation phrases, it would be rather unlikely that hedges were 
treated as separate IP’s unless they also express hesitation on the part of the speaker, as in (13):
(13)  Lionel: Look, there’ll be no competition. I mean, Margaret and clothes, 
 | they " just | " sort of | hang on her.
It is far more likely for hedges to be incorporated in the same intonation phrase as the part of the 
discourse they are referring to. As such they are either completely unstressed (14) or they may be 
part of the pre-nuclear segment (15), and sometimes even carry the nuclear tone (16):
(14)  Jean: She was very glamorous, though, until she started weaving about.
 Lionel: Well, that’s thanks to old Butterworth’s millions, I suppose. 
(15)  Jean: I sup#pose (some women | do get more attractive the older they get.
(16)  Lionel: Does that make any sense to you?
 Jean: Well, in a ( sort of way.

A number of adjectives, adverbs, phrases and even clauses are used to express the speaker’s attitude 
or stance towards the message. #e prosodic treatment of a stance marker largely depends on its 
position in the utterance, the context of interaction, but above all on the speaker’s perception of 
its importance for the ongoing discourse. 
In example (17) Jean is disturbed by the fact that Lionel’s ex-wife is so glamorous-looking 
although he has always claimed the opposite. Hence her emphatic prosodic treatment of the 
hedge honestly which she pronounces as a separate intonation phrase and with the fall-rise tone 
which is often used to express an implicature or a contrast – in this case, Lionel’s dishonesty. 
(17)  Jean: (&Honestly Lionel, | ages ago when I asked you about her, why didn’t 
 you say she was a very glamorous women? I could’ve lived with that. I 
 wouldn’t’ve liked it, but I could’ve lived with it.
If a stance marker is a clause, there are good chances that it will be treated as a separate intonation 
phrase, especially if it is used to introduce an important piece of information which is in contrast 
with the general idea, as in (18), or common knowledge, as in (19):
5
(18)  Margaret: I don’t want to feel better. I’m apologizing for my behaviour, not my
   life. Oh! !e  (& truth is | I wanted to show Lionel that I didn’t remain the dull, 
   boring wife you always thought I was.
(19)  Margaret: Well, | oddly e(&nough | I do remember what information I 
 managed to wring out of you last night. It’s a love story?
In both cases the contrastive meaning is highlighted by means of the fall-rise tone.
 Interjections are frequently used in speech if speakers want to express their emotional excitement 
about the message or when they want to warn the addressee about something. Due to this 
function, they are always treated as separate intonation phrases while the default tone is the fall, 
as in (20):
(20)  Margaret: #(Oh! | You’re always hungry! | #(Hey! | Why don’t we all make 
 an evening of it?
 Margaret, Lionel and Jean know that Margaret had a bit too much to drink the previous night.

Since interjections are emotional and attitudinal expressions, other pitch contours are also 
possible. In example (21) the interjection oh is pronounced with the rise-fall indicating Margaret’s 
amazement about the fact that Jean does not dye her hair but is happy with her natural colour:
(21)  Margaret: What’s your natural colour, Jean?
 Jean: #is.
 Margaret: &( Oh!
An additional attitudinal dimension is provided by the key or the pitch height at which the 
interjection is said. If an interjection is pronounced with the low key, it often sounds either 
very formal or unimpressed, as in example (22):
(22)  Margaret: But I don’t think I ever said I found you a very dull, boring husband.
 Lionel: $( Oh!
 In the paper I tried to shed some light on the prosodic features of pragmatic markers, i.e. those 
lexical items which are indispensable elements in speech but which at the same time stand out of 
the sentence or clause structure. I followed the division of pragmatic markers into discourse markers, 
hedges, stance markers and interjections, as suggested by Carter and McCarthy (2006). 
#e prosodic analysis which I carried out on a sample of the BBC TV series As Time Goes By has 
shown that the intonation of pragmatic markers is not very straightforward but very dependent 
on the speaker’s perception of how important a particular marker is. #ere are, however, some 
prosodic patterns which are more typical of discourse markers on the one hand, and other 
pragmatic markers, on the other.
Discourse markers whose function is to open or close a conversation or to mark transitions from 
one topic to another are most likely to be treated as separate intonation phrases. #e same goes 
for those markers whose function is to make reference to shared knowledge between the speaker 
and the listener. Stance markers and hedges are often incorporated in the same intonation phrase 
unless the speaker feels the need to highlight a marker for some particular reason. 
If a pragmatic marker is treated as a separate intonation phrase, it may be said in a variety of tones; 
the choice depends on the function as well as the meaning of the marker. #us the falling tone is 
expected for those discourse markers which function as openers, closers of topics or conversations, 
as well as when they function as focusing devices. For discourse markers whose function is to 
refer to the shared knowledge between the speaker and the listener, two or even three tones are 
possible; if the speaker assumes that the addressee shares the same information and hence does 

not expect a response, the falling tone will be used. If the speaker wants to elicit the addressee’s 
response in order to check the common knowledge, the rise or the fall-rise will be used.
Stance markers and hedges are treated as separate intonation phrases either when they express the 
speaker’s hesitation or when their contribution to the discourse is so important that they warrant 
special emphasis. In the former case they are often said with the level tone, whereas in the latter, 
they may be said either with the fall or the fall-rise. #e fall introduces a new point of view, 
whereas the fall-rise is likely to be used when the marker introduces a contrasting stance.
In conclusion I would like to stress the fact that there is some correlation between the intonation 
of pragmatic markers and their position and function in the discourse. Although the interface 
is rather arbitrary, there are, nonetheless, some general guidelines which speakers are likely to 
follow.
 , broadcast by BBC Prime, (recorded 8 November 2006).
Brazil, D. 1997.  Cambridge: Cambridge University Press.
Carter, R., and M. McCarthy. 2006. Cambridge: CUP.
Schiffrin, D. 1987.  Cambridge: CUP.
Stenström, A-B. 1994.  London: Longman.
Wells, J. C. 2006. Cambridge: CUP.