Look whos talking

Document Sample
Look whos talking Powered By Docstoc
					Look who‟s talking
Textual practices in the Hutton
             Encounters between
•   Cordis
•   Hutton
•   CADS
•   Priming
•   Critical appraisal

• Analyses using Wordsmith tools and a pilot
  version of Xaira
  Issues around reflexive language
• Attribution choices and responsibility
• Citation and scare quotes
• Plagiarism and the sound bite
• Modular organisation and a variety of modes:
                             Tokens Files
  Transcript mode
US congressional records 1,265,206 138
UK parliamentary records 778,502         62
WH White House briefings 511,474         99
HI Hutton inquiry transcripts 913,149 48
              Transcript data
• A number of dialogues going on
• Participants are constrained by:
• Need to remain coherent with previous
  statements and documents. (retrospective)
• All being recorded, transcribed and put on
  web sites (prospective)
• Recurrent patterns
• Number 10 = testimony of Blair, Campbell
  and Powell (39,304)
• GOV = testimony of all Government witnesses
• (124,464 tokens)
• All Hutton = (849,395)
               Textual practices
• The practices leading up to the events which caused
  the Hutton inquiry to be set up were issues of:
• Fuzzy attribution/endorsement
• Forced priming

•   Dossier – ownership and sexing up
•   Dodgy dossier
•   Gilligan‟s 6.07 broadcast
•   Naming of Dr Kelly
•   Gilligan‟s notebook
• Using both quantitative and qualitative data
• Moving between keywords and concordances
• Salience and context

• Not everything you can count counts (Einstein)
     “Other bits of language”
• extra-vocalisation involves the quoting or
  referencing of the statements or points of
  view of external sources. Writers/speakers
  also use extra-vocalisation to position
  themselves dialogistically with respect to
  actual and potential communicative
Corpus based study of reporting

• G. Thompson: Reporting (COBUILD)
• Semino and Short 2004
• Corpus Stylistics : Speech, writing and
  thought presentation in a corpus of English
• Systematic and detailed annotation of a
  corpus of written fictional and non fictional
  narratives 258,348 tokens from 120 texts
• First of all: What can be counted?

• “Quote unquote” – in transcripts
  punctuation gives it away and even the
  machine can do it
            Easily counted
• There are 3 123 direct quotations ( 87,378
• 9.234% of the text
• Problems of transcription

• Using XML tags
               Making distinctions
• <q>="David_Kelly"></q> this is used when someone's, in
  this case DK's, actual spoken words are being quoted
  and the attribution is made clear by the person being
  named                      (0.389%)

•    <soCalled>***</soCalled> this is used when one word
    or an expression is being taken out of context
    and possibly marked out by intonation as being a
    quotation, there is no attention paid to whether this is a
    quotation taken from someone else in particular and so
    there is no attribution    (0.036% )

•  <writing type="***">***</writing> this is used when the
  written word is being quoted and the only attribution is to
  the document being read from and available as part of
  the Inquiry docs.
•                           (8.809%)

• “Rearrange into a well known phrase or
  saying” – recognition, even from incomplete
  samples, is part of socio-cultural competence
• Scare quotes
• A question of priming

• A question of interpretation
• “Indirect reports signal more clearly the reason
  for reporting the speech. Speakers can
  manipulate the use of these forms in such a
  way as to achieve a variety of social ends.”
• John Lucy
    Reporting verbs are countable
• Searching the HI corpus for reporting verbs
  signalling speech or thought presentation
  (Semino and Short‟s list)

• GOV: 3.03% of total tokens All Hutton: 2.5%

And across corpora…..







White House
    Text nouns signalling intertextuality
              are countable
•   Discussion
•   Draft
•   Meeting
•   Report
•   Drafting
•   All are salient as keywords
             In the keywords

• We find:
• Reporting verbs for thought and speech
• Text nouns
• But also
• Patterns involved in dialogistic positioning
           ways of positioning
• “the arguability of any utterance can be varied
  by adjusting the dialogistic status of the
  utterance, to engage with past, present or
  future communicative exchanges, taking up,
  acknowledging, responding to, challenging or
  rejecting actual or imagined prior and future
• (Martin and White)
• Endorse - brings in other texts for support
• Concur - aligns with interlocutor
• Pronounce - intrudes self

Typically against some opposed alternative
• increases the interpersonal cost of any
  rejection/doubting/challenging of the author's
  dialogic position
• Dialogic opposition
• resources by which some prior utterance
  or some alternative position is invoked so
  as to be rejected, replaced or dismissed as
           Getting at patterns
• Keywords help suggest what the core of a
  particular pattern might be
• We see the keywords correspond to resources
  linked to a number of dialogistic patterns:
                  Number 10

• Blair Campbell and Powell

• The data, wordlists, keywords, concordances,
  point to a habit of dialogistic second guessing
  and pre-empting with a tendancy to contract
  the dialogistic space.
• Proclaiming and Assertion – “tis so”
• Denial – “tis not”
• Probabilising via privileged access
• Relativising or spin – “just look this way” –
  backgrounding by foregrounding
• Justification “Noblesse oblige” closing down
•   Assertion: in the reporting verbs
•   Existential and locative expressions
•   Cluster           number 10 All Hutton
•   this was          0.19 %       0.07%
•   that was          0.27 %       0.15 %
•   it was            0.63 %       0.07 %
•   there were        0.27 %       0.15 %
•   there is          0.21 %       0.09 %

• Tot:              1.57 %     0.53 %
• Not
• No
• Contradictions of actual or imagined positions
     Denial across the corpora
No                 Not
• When epistemic markers are used it is usually
  with privileged access, located in individual
  subjectivity :
• We felt/We thought/I think

• Privileged accesss preempting rebuttal
• Not open to contradiction
• Mental processes of others she felt
• Grammatial foregrounding
• Cleft and pseudo-cleft constructions

• Thematic equations
  What it was doing was
       Spin – angling the position
•   Lexical foregrounding in the keywords
•   Focussing: what the point is
•   emphasise/important/key/salient/vital/.
•   Stance adverbials in the keywords:
              “Look this way”
• Point as a keyword
• Dialogistic positioning with respect of
  interlocutors (the BBC’s point of view)
• Limiting the range of choice dialogistically
  (there is no point, little point, a key point, )
            Noblesse oblige
• Modality: deontic and epistemic, perceived
  needs and duties, perceptions of reality

•               Number 10        HUTTON

• Having to:     0.22%          0.04%
• No options but one…….
           Spin and attribution
• Attribution is obviously `heteroglossic' in that
  it introduces an additional voice into the text.
  Many of the textual practices of spin blur
• Lack of attribution in heteroglossic texts is
  perceived negatively in the case of both
  plagiarism and sound-bites.
           Priming in the lobby
•   Non attribuable statements off the record
•   Briefing against
•   20 questions
•   False attribution of statements, a key issue
    involving concealment of addresser, what
    might be called „false‟ footing
             Priming process
• “A team of briefers work with Blair to prepare
  for Prime Minister’s Question Time,they then
  provide briefing on the lines and phrases to
  any senior politician appearing on Question
  Time, Any Questions etc. Written briefings are
  sent out to all MP‟s so they know the phrases
  to use and the line to take.” (Clare Short)
• Priming
           Textual Practices
• on-message responses known as „singing from
  the same hymn-sheet‟ (a blurring of addresser,
  message and timing, and blurring of
• Flooding clusters into the discourse for forced
          Attribution – an issue

• Sound bites start as <q> but their attribution is
  hidden and they often become <SoCalled>
  People think attribution is an important issue,
  but not everything that counts can be counted
• What identifies the soundbites in a corpus?
          An example:Centrally

• A keyword
• Found virtually only in Hutton inquiry
• Found in government testimony only
• Found in a pattern repeated by five different
• All downplaying and denying the role of Dr
• “Not centrally involved in the preparation of
  the document”
            Textual practices
• Do show up differently in and across corpora
  but demand tagging, involving close reading to
  identify, and interpretation
• Can sometimes be counted but not always
• Often are the result of technologisation of
  discourse by communication experts

Shared By: