The annotation of discourse markers in an English-Hungarian parallel corpus:

parsing out the complex relationship between formal and functional features
1. Introduction 5. Some results of the quantitative analysis
Motivations: Cluster analysis of of course Distributions of the mean F0
Distributions of the mean durations of DMs of mondjuk (~say) among female speakers
there is an extremely complex relationship between the formal and functional
persze as a translation equivalent
of course
properties of DMs (cf. Aijmer 2013: 18); of course I / Im
of course not
40
24
the prosodic realization of DMs still doesn't receive enough prominence in DM of course it / its 25 15
right/all right/that's right
27
of course you / youre 17
research (Wichmann et al. 2010: 103); of course he 10
16 sure

there is general agreement in the literature that a contrastive analysis can help of course the 6
16 additive / contrastive
26
well of course 6
tease out the diversity of meaning relations that semantically bleached DMs mark (cf. No, of course 5 other DMs and DM clusters
Yeah, of course 5 (oh, well, you know, etc.)
Mortier & Degand 2009). ,

Objectives: LXS CON

to identify the most relevant formal features of of course that help us disambiguate Position of of course (ht, na, ) persze (~goes
without saying)

between textual and interpersonal as well as contrastive and non-contrastive 180

160 Translations of of course
2
termszetesen (~naturally)

2
functions; Utterance positions of of course Wichmann, 140
Simon-Vandenbergen, Aijmer (2010) 120 2 j/szp, hogy, mg szp
4
100 5 (~sure it is)
to use these formal features in order to model the basic discourse functions of of 80 Initial 6
Initial Medial Final total nan (~very inf. 'sure')
60 Medial
101 72 (36%) 27 200 40 11
course via a decision tree, which can be tested on a larger corpus. (50.5%) (13.5%) (100%) 20
0
Final
59 dehogy / dehogyis /
dehogyisnem (~on the
Utterance positions of of course in Wichmann
House Seasons 1-4 House S1-5 contrary)
et al. s (~and)

2. Research material: scripted dialogues

Initial Medial Final Total
Turn and utterance Initial after mainly, Turn 15 3 200 is (~too)
well, etc. internal (7.5%) (1.5%)
LAC: First four seasons of House (also known as House M. D., NBC Universal 135 (67,5%) 25 (12.5%) 22 (11%)
br (~although)

Television) transcripts rather than subtitles

LBC: Hungarian translations (2 versions) transcripts of dubbed releases

837,088 words -~5% (characters names and stage instructions) 6. Conclusions 1: A decision tree for the functional disambiguation
LAC+LBC1-2 aligned (SDL trados) between interpersonal and discourse uses of of course
200 tokens of of course (251 tokens per million words ~ 190 tpmws in NOSD)
Steps of the analysis:
Distinctive features of Interp. / Textual
3. The functional spectrum of of course of course in order 0.5 / 0.5
of importance/distinguishing power
Previous research on of course: (25 / 25 tokens)

Lewis (2006): (1) emphatic yes, (2) concession, (3) background, (4) end of list
1. host unit in an
Aijmer&Simon-Vandenbergen (2004): (1) evidential,(2) interpersonal,(3) indeterminate adjacency pair
Furk (2007): (1) conversation management, (2) information management, (3) story Second-pair part
Non-second-pair
structure, (4) interpersonal uses part host unit
2. prosody (integration) host unit (27)
(23)

Basic functions to be Sample utterances

disambiguated in the corpus Separate Integrated
Separate Integrated
Prosody (6) prosody Prosody
TEXT Prosody (19)
(21) (4) INT
INTERP (=interpersonal and Amy: I'm just a little nauseous I umm... I think I ate too
interactional function is more much. Can we take a break? 3. position in turn
salient than textual function)
60% of tokens Henry: Of course.
Turn-medial Turn initial
CONTR (=contrastive and Wilson: I'm not gonna date a patients daughter. Turn initial Turn-medial
or final (2) INT
concessive textual function (19) or final (17)
(2) TEXT
including reformulation) House: Very ethical. Of course, most married men would say
30% if tokens they dont date at all. 4. transl. term. / dehogy(is)
ADD (=additive/non-contrastive Let's see, your stomach has the deep-seated feelings of abandonment written all
over it which points towards sexual abuse. Well a fear of hospitals; that points to a
textual function including result, more specific traumatic event, so I'm going to say-- your mom, in the hospital with a
elaboration and justification) candlestick. And by candlestick, of course, I mean inherited OTC deficiency. Not transl. Not transl.
Term. / deh. as term/deh. Term./dehogy
10% of tokens as term/deh.
(16) (15) (2) INT
(3) TEXT
Marginal resultative House: Are you tired?
(interactional function is also Dan: Sometimes. 5. prosody (stress)
salient: SPP+int. pros.+mid-turn) Dad: He never sleeps! Of course hes tired.

Unstressed Stressed Unstressed Stressed

(1) TEXT (15) (12) (3) INT

Prototypical set of Prototypical set of

features of the features of the
INTERACTIONAL TEXTUAL
uses of of course uses of of course

7. Conclusions 2: A decision tree for the functional disambiguation

between contrastive and additive (non-contrastive) uses of of course
Steps of the analysis:
Distinctive features of
of course in order
of importance/distinguishing power CONTR / ADD
0.5 / 0.5
(25 / 25 tokens)

1. position in utterance

Initial Medial or final

4. Annotation tags CON (23) ADD (27)
2. stance
Formal properties:

Prosody (stress and integration)

Other than ir. Irony Other than irony Irony
Relevant co-text (DM clusters, collocations) ADD(4) CON (19) ADD (20) CON (7)
Speaker roles (doctor to doctor, doctor to patient, exchange bw friends, etc.)

Speech act of the host utterance / preceding utterance

3. translation (persze-default)
Position in the utterance (initial, medial, final)

Position in the turn (initial, medial, final)

The host unit's position in conversational structure (second pair part/reaction/loose Persze Other than
Other than
connection) persze (2) CON (17) persze
ADD (16) persze (4)

Functional properties:

Textual functions (contrast, concession, addition, result, elaboration, justification)

Stance (irony, boost, uncertainty)

The annotator's confidence in tagging the textual function of the DM Prototypical set of Prototypical set of
The annotator's confidence in tagging the speaker's stance
features of the features of the
contrastive non-contrastive
use of of course use of of course
References
Aijmer, K. 2013. Understanding Pragmatic MarkersA Variational Pragmatic Approach. Edinburgh : Edinburgh University Press.
Aijmer, K. Simon-Vandenbergen, A-M. 2004. A model and a methodology for the study of pragmatic markers: the semantic field of

expectation. Journal of Pragmatics 36: 17811805. Pter Blint Furk

Chovanec, J. 2011, Humour in quasi-conversations Constructing fun in online sports journalism. In: Dynel, M. (ed.) The Pragmatics of

Humour across Discourse Domains. Amsterdam/Philadephia: John Benjamins, 243264.

Kroli Gspr University of the Reformed Church in Hungary
Dynel, M. 2011. Stranger than fiction? - A few methodological notes on linguistic research in film dicourse. Brno Studies in English 37(1),

4161. furko.peter@gmail.com
Furk, B. P. 2007. The status of of course as a discourse marker. HUSSE 8 Conference Proceedings. http://husse-esse.hu/wp-

content/2007/04/furko-peter-of-course-as-a-dm.doc.
Furk, B. P. 2011. A Contrastive Study of English of course and Hungarian persze. Proceedings of the HUSSE 10 Conference, Linguistics

Volume. Hungarian Society for the Study of English, Debrecen.

House, Juliane. 2015. Translation studies and Pragmatics. Paper presented at IPrA 2015.
DiSpoL 2015
Lewis, D. 2006. Discourse markers in English: a discourse-pragmatic view. In Approaches to Discourse Particles, edited by Kerstin Fischer,

43-59. Amsterdam: Elsevier.

Saarland University, Saarbrcken
Mortier, L. and L. Degand. 2009. Adversative discourse markers in contrast. International Journal of Corpus Linguistics 14(3): 33966.

Wichmann, A., A-M. Simon-Vandenbergen, and K. Aijmer. 2010. How prosody reflects semantic change: A synchronic case study of of
1-2 October, 2015
course. In Subjectification, Intersubjectification and Grammaticalization, edited by Kristin Davidse, Lieven Vandelanotte and Hubert
Cuyckens, 103-155. Berlin and NY: De Gruyter Mouton.