Resources versus constraints and limitations in interpreting

1
Resources versus constraints and limitations in
interpreting

• Daniel Gile
• ESIT, Université Paris 3 Sorbonne-Nouvelle
• daniel.gile_at_yahoo.com
• www.cirinandgile.com
• www.est-translationstudies.com

2
Thanks to CIT

• Keynote with little knowledge of
  signed-language interpreting
• Embarrassing

3
Good vibes

• Excited at the prospect of learning a lot from
  you
• Met some of you in Salt Lake City earlier this
  year
• Impressed by the human vibes I felt
• In Salt Lake City earlier this year
• Here since I arrived!
• Your work is important
• and makes a difference in the life of many
  people.

4
Signed Language Interpreting

• Something to be proud of
• and training signed-language interpreters
• also important,
• you are devoting a lot of time, thought and work
  to the task
• Visited San Antonio college yesterday
• Every aspect of the learning environment has been
  carefully designed to make it efficient and
  comfortable
• (Lauri Metcalf)
• Creative thinking and method using technology
  (Tom Cox)

5
Topics in this talk

• Fairly general
• Based on experience of a
• (spoken language) conference interpreter and
  translator, trainer, researcher
• No experience or direct knowledge of public
  service interpreting
• Aware of some questions and issues
• discussed in SLI community
• Picked up a few ideas
• that I hope will be of some relevance

6
Relevant?

• No money-back guarantee !

7
One view of Translation/Interpreting (1)

• 1. A service
• has to be useful as such
• so needs to meet actual needs and expectations of
  users.
• 2. Provided with resources Language, cognitive
  and technical skills, a working environment, some
  time, pre-existing thematic knowledge, technical
  equipment etc.
• 3. Provided under constraints Limited time,
  interference from people and other action, you
  dont see the screen, limited cognitive
  resources, limited knowledge of the context,
  fatigue, expectations from principals, working
  norms etc.

8
One view of Translation/Interpreting (2)

• Translating/interpreting always involves
  interpretation
• (in the wide sense)
• in particular, trying to make sense of the
  sounds/signs/words/situation as perceived by the
  translator/interpreter
• Translating/interpreting always involves
  decisions
• - Decisions when trying to understand a
  speech/text e.g. when interpreting images or
  sounds as corresponding to words or signs or some
  other expression of meaning
• - Decisions when composing ones own response,
  essentially target speech/text. Choose words,
  sentence structure, the tactic to address a
  problem

9
A model of Translation/Interpreting

• Source Text
• Interpretation
• Resources
  Constraints
• Decision making
• Target Text

10
Translation/Interpreting Quality is not absolute

• What is the best service for a user?
• - Full and faithful conveying of information?
• - Clarification?
• - Simplification?
• - Explanation? Tutoring?
• (in educational settings)
• What if for same assignment, different
  expectations?
• (the accused, her/his counsel, the prosecutor,
  the judge)
• Is standard ASL the best for a deaf client who
  generally uses a regional variation of ASL?
• Is standard ASL terminology better than
  an ad-hoc sign in specific settings?

11
A Translator/Interpreters quality

• Can you expect the same quality from an
  interpreter who is given background information
  as from one who has to start interpreting without
  it?
• Does an interpreter understand a client who signs
  in a variation of ASL s/he is not familiar with
  as well as a client who signs in a variation
  known to her/him?
• Does an interpreter provide the same quality of
  service in fields s/he is familiar with as in
  fields s/he is not familiar with?
• Can an interpreter provide the same quality when
  s/he sees the speaker well/screen and when s/he
  does not?
• More generally high variability in the
  difficulty of speeches.
• partly unpredictable

12
The Interpreters Role the adequacy of the
Conference Interpreting Model (1)

• Much of the pioneering work in establishing
• - The professional status of interpreters
• - Interpreter training methods
• - The academic discipline of Interpreting Studies
  
• was done within Conference Interpreting
• Principles developed by conference interpreters
• are often viewed as a model
• But the Conference Interpreting model
• Is not necessarily the most appropriate for SLI
• Or for Public Service Interpreting
• Or for some other types of interpreting

13
The Interpreters Role the adequacy of the
Conference Interpreting Model (2)

• Conference Interpreting generally applies to
  situations where
• The principals in the communication process are
• - Educated
• - Socially at similar levels
• (though there can be power differentials)
• - Represent collective entities or teams
• - Are not facing personal crises/problems in the
  interpreted event
• In such situations, the neutral conduit model has
  justification
• What about other situations,
• Especially those encountered in SLI?

14
In other settings

• In SLI, there are many cases where the neutral
  conduit role
• is clearly not the best to serve the interests of
  the users
• Examples
• In educational settings,
• it is in the interest of both teachers and deaf
  students
• to have the interpreter not only transfer the
  information, but also help the students in other
  ways
• inter alia clarify, explain, even give moral
  support
• Some research in court interpreting and health
  care interpreting shows that some principals,
  including judges, consider some interventionism
  of the interpreter acceptable

15
Cooking ingredients for the interpreters role

• A tentative model of the cooking components
• 1. Core language-barrier related needs
• (non-controversial, because the principals
• do not understand each others language)
• Cultural barrier as well? Sometimes
• 2. Further situation-specific communication
  needs/norms
• 3. Further wishes of the parties
• (generally) more from clients and principals
• Less from interpreters
• Role ? Core Specif. Needs/Norms Wishes
• Relative strength of wishes a major determinant

16
The interpreters role results from a balance of
power?

• Can/should there be an absolute definition of
• the interpreters role and working conditions,
• applicable to all cases?
• They are determined to a large extent by
  negotiation between
• Clients, principals in the communication
  interaction, interpreters
• Historically speaking, conference interpreting
  started out in a strong position
• Conference Interpreters were in a position
• to determine their role themselves
• This is not the case of other types of
  interpreters,
• and in particular court interpreters
• and signed language interpreters

17
Can/should SLIs determine their own role(s)? (1)

• Complex issue mostly because components of the
  role are sometimes difficult to disentangle but
• Different attitudes towards different components
  of the role
• Core needs
• (associated with lack of linguistic
  communication)
• - They need to be understood and
  discussed/clarified with clients and principals
• - Instructors and professional bodies could have
  an important role to play
• Are there cultural core needs in SLI?
• Legal norms
• Cannot be negotiated
• though clarification with court officers may be
  required

18
Can/should SLIs determine their own role(s)? (2)

• Wishes
• Some can be considered abusive interpreters are
  interpreters, not secretaries, not reporters, not
  waiters
• Professional organizations may be able to do
  something to protect their members
• Some could be negotiated
• Interpreters may accept to take on the role of
  interpreters-cum-tutors, interpreters-cum
  advisors etc.
• but perhaps with further professional
  qualifications
• and for higher compensation?
• Such roles are beyond Core needs!
• Is it feasible to negotiate???

19
Role, Fidelity and Ethics

• Service quality should be assessed largely
• as a function of the local definition of the
  interpreters role
• (and of local working conditions)
• (in the particular situation, on the basis of
  previously accepted norms)
• This applies in particular to informational
  fidelity,
• i.e. to the degree of equivalence considered
  best between informational content of the source
  speech
• And informational content of the target speech
• (what information to add/to delete/to change)
• Also note, (will be discussed later), that
  maximum informational fidelity depends on
  available cognitive resources

20
Information and utterances general laws

• General laws for your consideration,
• which have implications on fidelity strategies
• 1. Generally, there is more information in an
  utterance
• than the information that the utterer wanted to
  convey
• 2. Generally, utterers are not fully aware
• of all the information carried by their
  utterances
• 3. Generally, in spoken languages (and in signed
  languages?), inter-speaker and intra-speaker
  variability is expected
• (inter-speaker variability different speakers
  tend to produce different utterances for the same
  idea
• intra-speaker variability it is not unusual for
  the same speaker to produce two utterances for
  the same idea at different times)

21
Evidence for the general laws (1)

• Situation
• In a conversation with you, a friend says he
  wants to know the date and place of birth of
  Woody Allen.
• You ask him to wait a second, switch on your
  computer and find the following on the internet
• Woody Allen was born December 1, 1935 in
  Brooklyn, New York
• You want to convey the information to your
  friend.
• What exactly would you say, in English or in ASL?

22
Evidence for the general laws (2)

• Probably nearly every speakers utterance will
  be different
• Try it out in your classes
• Other examples
• Clinton ? Bush ? Obama ? ?

23
Evidence for the general laws (3)

• In all experiments conducted so far
• (dozens, with a total of hundreds of participants
• and many languages)
• 1. Almost as many different utterances as there
  were utterers
• 2. In repeat utterances, frequent variability
• (the 2nd time, people did not say the same things
  as the 1st time)
• 3. People tended not to be fully aware of
  variability in their own utterances

24
Examples of inter-individual variability

• (1) Paris is 50 kilometers away
• (2) We are still 50 kilometers from
  Paris
• (3) 50 more kilometers

25
Informational differences general

• (1) Paris is 50 kilometers away
• (2) We are still 50 kilometers from Paris
• (3) 50 more kilometers to go
• (4) We will be in Paris in 50 kilometers
• (2) and (4) indicate there are at least two
  people involved
• (3) does not indicate the destination
• (1) does not indicate that Paris is on the
  speakers route
• (1) does not indicate a specific event
• (that the speaker and at least one person is
  moving toward Paris)

26
Informational differences Framing Information

• (1) Paris is 50 kilometers away
• (2) We are still 50 kilometers from Paris
• (3) 50 more kilometers to go
• (4) We will be in Paris in 50 kilometers
• Some information is selected
• Consciously or not
• To help the addressee understand the Message
• (the Message the information the utterer wishes
  to convey)
• If you just said fifty, the Message
• would be difficult to understand
• Unless
• the other person just asked How many more
  kilometers?

27
Informational differences Linguistically and
Culturally Induced Information (1)

• (1) Paris is 50 kilometers away
• (2) We are still 50 kilometers from Paris
• (3) 50 more kilometers to go
• (4) We will be in Paris in 50 kilometers
• Paris is present, singular
• We are present, plural
• In Japanese
• Indication of tense optional, personal pronouns
  optional, singular/plural indication generally
  absent
• But indication about social relationship between
  speaker and addressee mandatory
• Sometimes, mandatory indications about the
  speakers attitude and feelings
• Sometimes, mandatory indications about position
  of objects close to speaker or addressee or away

28
Informational differences Linguistically and
Culturally Induced Information (2)

• In signed languages
• Indication of gender? Of time?
• but relative positions in space, directionality
  of movement, etc.
• Sometimes, the language (or culture around
  language) requires us to provide some information
  
• beyond the information we want to convey,
• even in a simple statement which is only
  intended to provide a small amount of information
• LCII is not the same in different languages and
  cultures

29
LCII and fidelity (1)

• How should we deal with this LCII when
  interpreting?
• Failing to reproduce it unfaithful to the
  speaker?
• When you meet someone you do not know and greet
  that person
• Bonjour mademoiselle! ? Hi!
• Missing information
• - You are talking to a woman
• - That woman is not (yet) married
• So, is Hi! unfaithful?

30
LCII and fidelity (2)

• Should you translate the greeting as
• Hi, unmarried woman!
• ???
• Why not?
• Because the additional information is not only
  irrelevant,
• but also (potentially) damaging to the
  interaction

31
LCII and fidelity (2a)

• Hi, unmarried woman!
• ?

32
LCII and fidelity (3)

• In many situations where a speakers intention is
  to provide information, explain or convince the
  addressee, LCII are natural in one language
• but unnatural in the other
• If you introduce them in the other,
• you may distract the attention of the addressee
  from the actual message you want to get across,
• and even generate misunderstandings about the
  speakers intentions
• Therefore, as a rule, LCII need not be
  reformulated when interpreting

33
LCII and fidelity (4)

• Other situations may arise where a speaker wants
  the addressee to learn more about his/her
  language and culture
• In such a case, LCII is part of the message the
  exotic part

34
PI and fidelity (1)

• Signed-language interpreters report that when
  signing for the deaf, they tend to include
  information about the signer and his/her features
  as a speaker
• (pronunciation defect, regional accent, language
  errors)
• because the deaf person does not have access to
  it
• But
• Does this Personal Information help achieve the
  speakers objectives?
• Could it not distract the attention of the deaf
  person from the actual Message?
• Could it not actually produce an unfavourable
  impression of the speaker and thus reduce the
  impact of his/her utterance?

35
PI and fidelity (2)

• When asked whether they also tell the hearing
  person about a signers features as a speaker,
  many SLIs say no.
• Why not? The hearing person does not have access
  to such information, does s/he?
• By the way,
• I never heard a spoken language interpreter say
  s/he reports to each of the principals in an
  interpreted dialogue about the features of the
  other.
• Why not? What is the difference in terms of
  access to such information between two principals
  using different spoken languages and two
  principals, one who uses a signed language, and
  the other a spoken language?

36
Framing Information and fidelity

• LCII is imposed by the language of expression
• No special reason to keep it in the target speech
• Framing Information is selected
• to help understand the Message
• If it is useful in the target speech, keep it
• If it is not, discard it
• (both LCII and FI may have to be added in the
  target speech as well)

37
Cognitive constraints fundamental ideas (1)

• Speech production
• Speech comprehension
• (both spoken and signed)
• require much processing in the brain
• At any time,
• the brain has finite resources for processing
• Sometimes, competition from other activities
• takes away part of these resources
• and speech production/comprehension suffers

38
Cognitive constraints fundamental ideas (2)

• Such competition can come from
• Trying to make sense of a situation
• with insufficient background knowledge,
• From decision making
• What should I do/say? What is appropriate?

39
Cognitive constraints fundamental ideas (3)

• Potential effects of insufficient processing
  capacity
• on speech production
• - Speech production becomes slower
• (you need more time to access the words and rules
• and to combine them into utterances)
• - Speech quality suffers
• Diction, pronunciation, clarity of signing (?),
  grammar, style, appropriateness of words
• In speech comprehension, insufficient processing
  capacity leads to non-comprehension
• Ex. You know the signs, but at a certain speed,
• You do not understand the utterance at all

40
Cognitive constraints fundamental ideas (4)

• Potential effects of insufficient processing
  capacity
• on speech comprehension
• - Words are missed though you heard them/saw
  them
• - The utterance does not make sense to you

41
SPEAKER
HIGH AVAILABILITY
LOW AVAILABILITY
WM span
time
t1 t2 t3

At t1, high availability listener (HAL) has
finished processing more than 2 words and keeps
one in WM low availability listener (LAL) has
finished processing 1 word At t2, speaker is
uttering 7th word, HAL has finished processing 6
words LAL has finished processing 2 words, and
must keep 5 words in WM. At t3, LAL is probably
saturated
42
Cognitive constraints fundamental ideas (5)

• Speech production and comprehension
• are highly dependent on background knowledge
• In particular,
• you require less processing capacity (and less
  time)
• to understand an utterance
• if you know about the people, the context, the
  stakes involved in a situation
• You require more processing capacity (and time)
• If you do not have this knowledge

43
Cognitive constraints the Tightrope Hypothesis
(1)

• In interpreting speaking and listening at the
  same time
• Actually Speaking Listening Short-term
  Memory
• (Effort Model of simultaneous interpreting)
• So you need more processing capacity than if you
  only listened or if you only spoke
• plus you need to make sense of the situation
• plus you need to take decisions
• The Tightrope Hypothesis says that as a result,
• you tend to work close to saturation, i.e.
• Close to a situation where you may run out of
  processing capacity at various moments

44
Cognitive constraints the Tightrope Hypothesis
(2)

• The Tightrope Hypothesis offers an explanation
  for
• - most errors and omissions encountered in the
  field
• - many infelicities encountered in the field
• In order to reduce the number of e/os and
  infelicities
• it would make sense to
• reduce proc. capacity requirements of
  interpreting tasks
• starting with
• the requirements of
• speech comprehension and speech production

45
Reducing processing capacity requirements

• One way of reducing such requirements
• is to acquire background knowledge
• - about the needs of clients,
• - about professional situations
• - about behaviour norms (best practices)
• to prepare for each interpreted event as
  thoroughly as possible
• But this is not enough
• Language availability is paramount

46
Language availability (1)

• Setting aside context-dependent variations in
  processing capacity requirements for
  interpreting,
• one major determinant of such requirements
• is language availability
• In a nutshell
• You may know various components of a language
  and how to use them
• But using them for comprehension and production
• May require more or less processing capacity and
  time, depending on their availability

47
Language availability (2)

• Language availability
• Some kind of basic attribute of each language
  component you know
• (word, rule of grammar, of style, of pragmatic
  use, spelling, sign, pronunciation)
• With respect to the time and effort (processing
  capacity) you will need
• to use it for effective comprehension/production
• (some kind of fundamental level
• that can change with circumstances)

48
THE GRAVITATIONAL MODEL OF LINGUISTIC
AVAILABILITY (1)

• 
  Relative availability
• The closer to the center,
• the less time and effort
• required for
• production/comprehension

49
THE GRAVITATIONAL MODEL OF LANGUAGE AVAILABILITY
(2)

• 
  DYNAMIC, NOT STATIC
• 
  EVERY LANGUAGE COMPONENT
• 
  
• 1. DRIFTS OUTWARDS
• IF NOT USED
• 2. GOES INWARDS IF USED
• 3. ESCORT EFFECT
• 4. INTERFERENCE EFFECT
• 5. WRITTEN AND ORAL
  SYSTEM NOT IDENTICAL
• 
  6. PROD. AND COMPREHENSION
  
• 
  NOT IDENTICAL

50
Language availability (3)

• On the basis of these general rules
• High language availability in ones working
  languages
• requires repeated, frequent use of the relevant
  language components
• In production, evident
• In comprehension, note that
• frequent encounters with various accents
• (and personal/regional variations of signing)
• are necessary for good comprehension
• (this lowers processing capacity requirements for
  decoding the sounds/visual signals)

51
Language availability (4)

• In spoken language interpreting
• Language availability can be
• sufficient for consecutive interpreting
• and insufficient for simultaneous interpreting
• Many conference interpreting students fail in
  training programmes because of insufficient
  availability
• In signed-language interpreting
• Nearly always simultaneous (?)
• How do you manage?
• Perhaps OK for everyday interaction
• More problematic for non-routine topics?

52
The Tightrope Hypothesis and saturation (1)

• Since interpreters work close to saturation,
• Even when language availability is high
• Processing capacity saturation can occur
• - When encountering utterances with
• high general information density
• - High speed delivery
• - Read speeches
• - When high local information density of
  utterances
• speed of speaking/signing
• compound names, enumerations, finger spelling,
  unexpected information
• homophones or quasi-homophones
• (or signs that look similar)

53
The Tightrope Hypothesis and saturation (2)

• - When noise
• (discriminating signal from noise takes up
  processing capacity)
• - When unexpected information, word, rationale,
  idea
• When non-standard signing seen or required
• (a students wants you to use a certain sign you
  are not used to for a certain concept)
• - When tactical errors in interpreting
• (e.g. wrong decision which leads to a delay)
• Other cases
• Explained inter alia in Gile (2009). Basic
  Concepts and Models for Interpreter Training.
  Amsterdam/Philadelphia John Benjamins

54
CONCLUSIONS (1)

• 1.Perfect, error-free interpreting products
  a myth
• Errors, omissions and infelicities are so
  frequent that
• they can be considered a regular part of the
  product
• 2. The quality of the interpreting product
  depends to a large extent on constraints and
  resources
• 3. Some of the constraints and resources depend
  on the working environment and on the clients
  willingness to optimize conditions
• Employers should be responsible for working
  environment
• But interpreters should raise their awareness

55
CONCLUSIONS (2)

• 4. Professional organizations should explain to
  clients (and to practitioners, and to trainees
  through trainers)
• That some expectations are OK
• and some are not,
• Because they make it difficult to perform well
• Because they go against fundamental norms
• Because they go beyond an interpreters role
• And insist that in order to optimize interpreting
  services, principals should be active
  participants
• Not passive beneficiaries of these services

56
CONCLUSIONS (3)

• 5. Interpreter Trainers can maximize some of the
  resources, in particular by working on
• Language skills
• Coping tactics
• Strategies
• (preparation, further training,
  awareness-raising)
• 6. Researchers can contribute, especially by
  investigating the effect of constraints and
  resources on the process and product of
  interpreting