You are on page 1of 26

The Use of Modern Technology in Translation*

__________________ _Tayebeh Mosavi Miangah"

Abstract:

Rapid development o f translation industry in the world today

reveals the growing need for translation tools. T h is article tries to

briefly study various translation tools. These tools are considered

as very helpful devices in the translation process. T h e most

important tools involved are translation memory, machine-

readable dictionaries and machine translation softw are. Using

modern technology in translation not only accelerates the

translation process itself but also augments its consistency and

efficiency.

Key Words:

Machine translation, translation memory, m achine-readable


dictionaries, natural language processing system , pre-editing,

post-editing, machine-readable language corpus

’ This paper was received on 12/09/2004 and approved on 15/11/2004.


1 Assistant Professor at the English Department, Shahr-e Kord University;
e-fna»: mQusavi-iaiit.sku.ac.ir
Anticipation Strategy and Simultaneous
Interpreting1

___________________________ Karim Shabani”

A b s tr a c t:
A n ticipa tio n as an on -lin e strate gy in S I Is conceived as rendering
a co n stitu e n t in th e o u tp u t language b y th e sim ultaneous
in te rp re te r b e fore receiving its o rig in a l counterpart in th e in p u t
language. A p o tp o u rri o f in trig u in g cues, though piecem eal, is
n o rm a lly taken as a backdrop b y th e in te rp re te r to hypothesize on
th e speaker's u n fin ished utterances. The pre se nt paper takes its
p o in t o f de pa rture fro m d ie b a s t assum ption th a t anticipation is
reg ard ed as in evita ble in S I process p a rtic u la rly in th e case o f tw o
s tru c tu ra lly asym m etrical language-pairs. W ith th is in m ind, we
w ill, fo llo w in g an in trod uctio n, pre se nt an overview o f th e how
a n d w hy o f th e d iffe re n t types o f know ledge a t th e interpreter's
disposal a t th e tim e o f a n ticip a tin g during S I and then proceed
w ith m u ltip le exam ples taken fro m d iffe re n t live corpus to J ustify
th e m ainstream p o sitio n . F inally, on im plication side, the paper
w ill o ffe r som e tip s on how to fo s te r th e sa id strate gy fo r m aking
an a p p ro pria te an ticip ation.

Key Words;
Anticipation, simultaneous interpreting, asymmetrical, Speaker,
extra-linguistic

' This paper was received on 28/11/2004 and approved on 14/02/2005.


* Faculty Member at the English Department, University of Maritime Sciences,
10 T ransla tio n M s s , Vol. 2, No. 7 & 8, Autumn 2004 8t W inter 2005

In tro d u c tio n
Anticipation in simultaneous interpreting Is generally conceived as rendering a

constituent In the output language before receiving its corresponding

constituent in the Input language. During simultaneous Interpreting such factors

as the synchronitity of input and output, imm ediacy o f response due to the

short EV5 or time lag and thinking among others make anticipation more risky

or difficult. What doubles this difficulty Is som etim es th e language-specific

features of the language pairs. For instance, the asym m etrical w ord order or NP

structure of such language pairs as Perslan-English, Chinese-English and

German-Engllsh is assumed to cause a problem in SI. T o coun ter such a

problem, anticipation strategy is usually employed. (See also Gile: 1992, Setton:

1999, Shabani: 2001)

Some researchers like Wilss (1978) believe th a t anticipation in

simultaneous interpreting is considered as a language-specific phenomenon.

They claim that In the case of two structurally sim ilar lang u ag es, th e interpreter

doesn't need to embark on It for s/he can translate th e co n stitu e n ts almost

sequentially or when interpreting from an SO V lang u age to a n S V O language

like Persian and English, respectively, the interpreter c a n k ee p th e English verb

in his working memory until he receives th e n e x t co n stitu e n ts and then

produces it in his output when he sees no structural co n stra in t. Therefore, the

direction of interpreting can somehow affect th e d eg ree to w h ic h anticipation is

used. (For more on 'direction of interpreting' refer to M ah m o u d za d e h , 1994) For

some other authors, anticipation is the indispensable c o m p o n e n t o f SI task

irrespective of the structural asymmetry. T he y claim th a t a g o o d interpreter

should always predict the speaker's utterances a n d h y p o th e size on the

speaker's unpronounced messages in advance since w a itin g fo r s o long can

saturate him cognitively and subsequently h e w ill fa ll b eh in d th e speaker.

(Lederer, 1981)

Nonetheless, for Setten and IP-oriented re se arch e rs w h e n interpreting


Anticipation Strategy and Simultaneous Interpreting 11

from an SOV language like German, Chinese, Persian, Turkish or Japanese into
an SVO or VSO language like English and French, there is the problem of the
verb that Is needed early in output language but produced late in the source
language. (Setton: 1999) To counter these language-specific asymmetrical
structures or linguistic straitjackets, anticipation and three other strategies
namely chunking, stalling and waiting (Shabanl: 2004) are essential and
normally used by the simultaneous interpreters for producing a smooth and
continuous translation. T o expand on the topic, a further crystallization of the
first strategy i.e. anticipation as the cornerstone of SI, its components and
different levels at which it occurs was deemed essential. In this paper, the
author attempts to present the aforementioned crystallization in association
with the types of the cues available to support a plausible anticipation through
giving some evidences o f live corpora taken from different language pairs In SI

process. Since anticipation is made during the perceptive (listening) phase of

SI, th e following discussion Is mainly presented in association with the listening

process.

Nature o f Anticipation
A t the outset, the question arises as to how and when anticipation occurs.

Outwardly, there should be some basis to rely on so that one can carry out

prediction when processing the language. If we look at the languages of the

world and attem pt to classify them, we find both diversity - the differing order

o f Subject, O bject and Verb - and fundamental commonalities: ail languages

have verbs; all languages have N Ps that may function as Subjects and Objects;

and all languages have rules that ''linearize" these constituents into basic order.

N o language random ly orders its constituents. (Fromkin e t a!., 1983) These

specific characteristics that define the grammar o f a typical language can be

taken as an authoritative source for anticipation proper to be materialized.

Based o n the schem a theory, the human's mind is able to make hypothesis

about som ething unknown (new information) with th e help o f something known
12 T ransla tio n S tudie s, Vof. 2 , No. 7 & 8, Autum n 2004 & W in te r 2 0 0 5

(old information). This kind o f movement from know n to u n k n o w n h a p pe n s by

virtue of the regularity that exists at different /ayers o f la n g u a g e w ith o u t w hich

anticipation would be next to impossrWe. Vu/e (1996:4-5) n o te s th at:

people tend to behave in fairly regular w a ys w h e n It c o m e s to using


language. Some o f the regularity d erives fro m th e fa c t th a t p e o p le are
members of social groups and follow g e n e ra l p a tte rn s o f b eh avio r
expected within the group.

What fo ttw s is a m ore analysis o f the p red ictab le n a tu re o f la n g u a g e at

different levels I.e. syntax, semantics and p rag m atics th a t fe e d th e hu m an 's

mind to make inferences.

Anticipation is simply defined as the listen er's n o rm a l re a ctio n to the

linguistic and extralinguistic ambiguities throughout liste n in g th a t is a ssu m e d to

take place at different levels. As for the language fu n c tio n in g a t m u ltip le levels

namely; syntax, semantics and pragmatics, an ticipa tion a ls o s o u n d s to operate

at these levels. Each level has its own rules and co nstra in s.

At syntactic level, the Persian listeners o f E n g lish c a n e a s ily p r e d ir t the

incoming object and then verb o f the Persian in p u t se n te n c e a fte r h e a rin g the

subject by merely benefiting from his/her linguistic k n o w le d g e o f th e S O V word

order of Persian. This type of anticipation is b acked b y th e h u m a n speech

parsing mechanism (HSPM). To elaborate o n H SPM , G a rm a n (1990:313)

maintains that:

TTie essence o f parsing mechanism is th at th e lis t e n e r is w o rk in g with


Input that consists of lexically identifiable e le m e n ts , tem po rarily
sequenced; and is working towards d eterm in in g t h e re la tio n s h ip s of
these elements as a part of understanding p roce ss.

Kimball (1975) gives us a com prehensive a c c o u n t o f t h e H SP M . He

proposes some principles of surface structure p roce ssin g w h ic h a r e g iv e n below

in brief:

I. Parsing is top-down: an interpretation b eg ins fro m th e s e n te n c e node


Anticipation Strategy and Simultaneous Interpreting 13

and predicts the constituent elements.


2. Right association: words are attached to the nearest logical place on the
right.
3. New words: these are signalled by the grammatical function words.
4. Principle o f closure: phrases are closed down as soon as possible.
5. Processing: when a phrase has been parsed it is passed from short-term
memory into a stage o f deeper semantic processing.

Researchers like Richards (1983) propose that listeners use strategies


based upon the surface structure o f the sentence to unravel the syntax. Part of
this syntactic knowledge in English can be the Subject-Verb-Object sequence or
the noun phrase structure. Such sim ple strategies (SOV, NP decoding) are tried

first and if they don’t succeed then further parsing needs to be done.

In addition to syntactic knowledge, semantic knowledge commonly

referred to as knowledge of the relations between linguistic expressions such as

the words o f a language and their referents such as persons, things and events

in th e world provides th e interpreter with ample information to reduce the

cognitive processing load during SI. T he sem antic knowledge indudes the

knowledge o f collocations, idioms and proverbs among others. Drawing on the

"Longm an D ic tio n a ry o f A p p lie d L in g u is tic collocation is defined as the way In


which w ords are used together regularly and the restrictions on how words can

be used together, fo r exam ple which prepositions are used with particular

verbs, o r w hich verbs and nouns are used together.

T he following are some collocations in English:


1- p e rfo rm a n a ctio n
2 - m ake a n excuse
3 - d o a lo t o f dam age
4- m ake tro u b le
5 - m ake a m istake
6- ta ke a trip
7- p a y a tte n tio n
14 Transla tio n S tudie s, Vol. 2 , No. 7 ft 8, Autumn 2004 & Winter 2005

in the above examples, the verbs can be guessed or, at feast, their
cognitive processing effort to understand the whole phrase would decrease
once the object or one element is received. The elements to be guessed are
usually redundant and semantically neutral. The following constructions In
Persian are more examples of such verb and noun collocations that Include

redundant elements:
8- jamzadan (to sw eep)
9■ zamin khordan (to m i)
10-de! be daria zadan (to ta ke a ris k )
u-azdastdadan (to lose)
12'gut zadan (to deceive)
13 -elam kardan (to announce)
14-zendegi kardan (to liv e )
15-be donya amadan (to b e b o rn )
16■mored-e barresigharar dadan (to analyze)
17-ta fs ir kardan (to com m ent)
18- khesarat vared kardan (to In flic t a dam age)
19• be molaghat-e kasi raftan (to v is it som ebody)
20-be ettefa-e ka siresandan (to in fo rm som ebody o f)
21-goftegoo-ye bein-e tam m adonha (dialo gu e a m on g civiliz a tio n s)
22‘ d irya zo o d (so o n e r o r la te r)
23-kam -o-bish (m o re o r le ss)
24-nazm-o-ghanoon (la w a n d o rd e r)
25'bim -o-om id (hope a n d fe a r)

In addition to collocations, idioms and proverbs being part o f semantic

knowledge can also function as single units and be, somehow, predictable like

collocations.

Here are some examples:

Id io m s P ro v e rb s
26- add insult to in ju ry * A ctions speak lo u d e r th a n words.
Anticipation Strategy and Simultaneous Interpreting 15

2 7 -a d d fu e l to fire * W hite th e c a t's aw ay, th e m k e w ill play.


28 - m ake a h e y w hile th e sun shines * Too m any cooks s p o il th e broth.
29- b y h o ok o r b y c ro o k * The grass is alw ays gre en er
30■ b e m ore C atholic th a n pope on th e o th e r sid e o f th e fence.

Examples in Persian:

Idioms {English translations)


3 1 -D eia m barayash k h e ili ta n g shodeh. ( I m iss h im a lo t.)
32- D e l be darya zadam . ( I to o k a ris k .)
33- KaUash b o o -ye gh orm e sabzi m ideh. (H e has dangerous p o litic a l view s.)
3 4 -A zk a h kooh sakht. (H e m ade a m ountain o u t o f a m o le h t/t)
35- D o p a dasht, dopa g h a rz g e re ft (H e show ed a d e a r p a ir o f he els.)

(Moilanazar, pp. 50-53)

T he idiom s and proverbs are usually regarded as set phrases whose

m eanings are understood as a whole. Once some parts of the phrase is

decoded, the other parts can be recalled more easily and the problem of
semantic re-arrangement of individual lexical elements to come up with an

understanding of the whole phrase is obviated or psycho-llnguistically speaking,


the total required processing capacity (PRPC) o f the phrase would decrease due

to the presence of predictable elements and, therefore, the TRPC would not
saturate the total available processing capacity (TAPC). (cf. Gile: 1999)

Apart from the syntactic and semantic knowledge, there Is still another
authoritative and more reliable knowledge that makes anticipation more
comfortable and less risky. Pragmatics simply defined as the knowledge of
language use and how it is interpreted in situational contexts in association with
its users can be o f great help for the interpreter. It consists o f the knowledge o f
the world, speech acts, speech events, settings, context or situation, which is
formed as a set o f schemata in the human's mind to process and decipher the
meaning o f linguistic input.
16 Transla tio n Studie s, Vol. 2, No. 7 & 8, Autumn 2004 & W inter 2005

YuJe (1996:85) highlights the role o f schem atic or background

knowledge and maintains:


Our ability to arrive at interpretation o f the unwritten and the unsay
must be based on preexisting knowledge structure. These structures
function Jike familiar patterns from past experience that we use to
interpret new experiences. The most general term for a pattern of this
type is a schema. A schema is a pre-existing knowledge structure m
memory.

Quite conterminously, Widdowson (1981) elaborates o n the quality of

'schema'and defines it as:

Cognitive constructs which allow for the organization o f information in


long-term memory and which provide a basis fo r prediction. They are
kinds of stereotypic images that we m ap on to actuality in order to
make sense of it; and to provide it with a co here nt pattern. (Cited in
Zohrevandi, 1994:150)

In this line Anderson and Pearson (1988) identify tw o types o f schemata

essential to comprehension, te xtu al and c o n te n t o r 'scrip ts'. Textu al schemata

are related to the discourse level of conventions o f certain types o f interactions

e.g. the conventions of a meeting with an acad e m ic supervisor. Scripts,

according to Schank and Abelson (1977), are a special ty p e o f schema and

refer to our situation specific knowledge o f routine eve nts, their goals and

encounters. For example, a 'going to a football m a tch ' scrip t would have

¡nfomiaboft about the roles and actions o f th e cha racte r, places and the

sequence of events. Based on the relevance o f facts in th e scripts, reception of

one subscript can be reminiscent of its related su bscrip ts and, then, their

umbrella script (topic under discussion) w ill be know n q u ite autom atically.

On the significance of the role o f schem atic kn o w le d g e to process the

message, Oiler (1979, cited in Nunan, 1999:202) had e a rlie r entertained the

same view in saying that

The mental frameworks are critically im p ortant in helping us predict


Anticipation Strategy and Simultaneous Interpreting 17

and then cope with the exigencies of everyday life and without these
schemata, nothing in life would be predictable and if nothing were
predictable it would be impossible to function. The world would appear
chaotic.

There is ample research on the vital role of background knowledge to


Interpret a message. On the basis of schema theory, Alderson (1980) has
Identified three functions of background knowledge or schemata. First,
schemata provide the basis for filling the gaps in a text: no message is ever
completely explicit and schemata permit a coherent interpretation through
inferential elaboration. Second, schemata constrain a person's interpretation of
an ambiguous message. Third, it is by establishing a correspondence between
things known, as represented by schemata, and the givens in a message that

listeners monitor their comprehension and know whether they have understood
the message. This meta-comprehension function is extremely important in the
full understanding of the message.

Also noteworthy is the fact that understanding a message needs the

'relevant1 background knowledge to be present without which the text couldn't

be interpreted and would sound as a set o f purely linguistic forms with no

meaning associated with it. 8 ransford e t <?/. (1972:720) takes this line and
observes that

Even children's stories require many assumptions to link sentences


together to form coherent events. If the people lack the knowledge
necessary to make appropriate assumptions or inferences, most
passages would presumably seem incomprehensible.

The bottom line here Is that the presence of assumptions is as essential

as the activation of relevant assumptions and anticipatory ability on the part of

the interpreter as a listener Is itself fed by the degree to which he actively

makes ‘p re cisé as opposed t o ' im pre cisé types o f assumptions while processing

the newly received linguistic strings. In fact, right prediction is a function o f the

degree to which he spontaneously activates knowledge that is semantically


18 T ransla tio n S tudie s, Vol. 2, No. 7 & 8, Autumn 2004 & W inter 200S

consistent and congruous with to-be-heard information and the listeners must

activate knowledge that clarifies the significance o f to-be-heard information

(Alderson, 1980)

An introspective study by Brown (1990) differentiated the strategies used

by effective and Ineffective listeners, which also seem s to be true of

Interpreters. They found that effective listeners regularly used three strategies:
t. Self*monitoring, whereby listeners were checking their knowledge and
their output while it was taking place.
2. Elaboration, relating new information to previous knowledge and to any
new information.
3. Inferendng, guessing strategies to fill in gaps in comprehension caused
by insufficient linguistic knowledge.

Inferendng ability is rooted in the hum an's W M (working metnoiy)

capacity that is able to make predictive inferences due to th e presence of

causal sufficiency in the text events. (Linderholm, 2002) For instance, when

hearing that a delicate porcelain vase was just throw n against a surface, nearly

all listeners would predict that the vase would break as a resu lt (van den Broek:

1990). According to Linderholm (ibid), this capacity is m uch less in low WM

capacity readers/listeners since the generation of predictive inferences is

contingent on the amount of information allocated to predictive concepts In

LTM (long-term memory) as a function of th e WM capacity of the

reader/listener.

Willis (1981:134) lists a series o f m icro-skills of listening, which he calls

en ab lin g s k ills . They are:


* predicting what the people are going to talk about

* guessing at unknown words or phrases w ithout panic

* Using one's knowledge o f the subject to help o ne understand

* identifying relevant point; rejecting irrelevant inform ation

* retaining relevant points (note-taking)

* recogncing discourse markers e.g., w e ll; oh , a n o th e r p o in t Is;


Anticipation Strategy and Simultaneous Interpreting 19

fin a lly ; etc.


* recognizing cohesive devices, e.g. such as and which
* understanding different intonation patterns and uses of stress, etc.,
which give dues to the meaning and social setting.
* understanding inferred information, e.g. speaker's attitude and
intention.

Nicholson (1987b) gives an example of a wrong translation by a trainee


interpreter that w as caused by a lack of relevant schematic knowledge and
unfamiliarity with the topic. She described that the student translator produced
'decreased' instead o f 'increased' somewhere in his interpreting and stopped
short o f taking the advantage o f the present contextual clues in his input.
According to Nicholson, this wrong translation could be avoided with reference

to the pre-existing contextual information i.e. the speaker’s main line of


reasoning, his tone o f voice, his previous statements and discussions. (My

translation; Tahourian, 1995)

The following example in Persian can clarify the predictive power of


schematic knowledge. Suppose a speaker is delivering a speech on the issue of

' inflation' in a simultaneously Interpreted conference and produces this

sentence:
1 .slJL .Ulj*- tX<T Jl* o U u i i ? O t y « ; u -y l JU,

The Persian interpreter can easily make a plausible projection of the

following predicate and guess the speaker's intended verb through activating

the relevant background experiences and subject knowledge of the topic under

discussion i.e. In fla tio n . Since inflation entails that the money loses its value and

doesn't gain It, the deduction that the value of Rial will 'decrease' rather than

'Increase' as his understanding o f the speaker's way of concluding is made and

produced as the output. This is also true o f the following sentence whose two

verbs are predictable due to the presence o f pre-existing co-text and discourse

markers in the thematic part.


20 Translation Studies, Vol. 2, No. 7 & 8, Autum n 2004 & W inter 2005

* Xl¿Jjj¿¡l^s^l J j & JU-J> J j j fjy Jr1-^

«m ^i>i 0 o.Tj».

it goes without saying that the sim ultaneous Interpreter a s opposed to a

mere listener or even consecutive Interpreter Is m o re in n e ed o f th e contextual

knowledge, This pressing need emanates from th e insu fficien t linguistic input or

co-text. Thus, s/he would rather activate the relevan t sch e m ata before the start

of meeting through studying the related jargon list an d p re v io u s lectures on the

same topic so that he can refer to them to resolve th e p ro b ab le ambiguities

through making appropriate hypotheses o r p red iction s in a c o h e re n t form.

Another point not to be skipped Is th at in so m e o cca sio n s, the word

prediction and recognition can occur a t syllable b o u n d a ries. W h ile listening, 3

listener can guess the following syllables o f a m u lti-sy lla b ic w o rd a fte r decoding

the first or second part. For instance, in the fo llo w in g e x a m p le s th e w h o le word

can be easily guessed once the first syllables a re re ce iv e d a n d o f course not

isolated but In a contextualized form.

1 in.tro.duc.tion

2 au.to.ma.ti.caLly

3 si.tu.a.tion

4 a.tten.Hon

5 a.partments

6 pho.to,graph

7 d.vi.li.za.tlon

8 or.ga.ni.za.tion

Therefore, the simultaneous interpreter w h o is u n d e r th e tim e pressure

and short EVS would be better o ff relying o n h is a n tic ip a to ry a b ility to make a

guess and not wait for so long so that he can b u y e v e n th e m illiseco n ds. After

prediction, his job is just to check and m o n ito r w h a t h e h a s co n je ctu re d while

receiving the next chunks o f information.


Anticipation Strategy and Simultaneous Interpreting 21

T o co nfirm a n d recap o n th e p o in ts raised s o far, C h e rn o v (1 9 9 4 :1 4 ? )

in trod u ce s a m od el fo r th e Interpreter's p re d ic tio n a t th e tim e o f sim u lta n e o u s

Interpreting, w h ic h he c a lls "p rob a b ility p re d ic tio n m a ch in e ry ". H e cla im s th a t

th is m ach in e ry w o rk s in fo ur tiers, w h ich a re as follow s:

(a) so u n d p a tte m s(sy lla b le s e n co d in g p ho ne m es, in to n a tio n stre ss a n d o th e r


p roso d ic fe atures)

(b) g ram m a tical (sy n tactical an d categ o rical sem an tic fe a tu re s)

(c) sem an tic tie r p e r se

(d) se n se tie r p e r s e

H e g o e s as fa r as to p rese n t a ta b le for th is m a ch in e ry (m o d e l) b ase d o n

red u nd an cy (Inform ation co ntent) fro m th e level o f th e s y lla b le to th a t o f th e

w o rd , p hrase, u tte rance , co m m u n icatio n (te xt) a n d situ atio nal context.

A cco rd in g to C h e rn o v (1994: ib id ), th e s e le v e ls a re d istrib u te d am ong

th e tie rs a s follow s:

Redundancy level Tier

Syllable, word, phrase utterance (a) prosodic: sound patterns

Phrase, utterance (b) syntactic: grammatical features and


semantic categories

Phrase, utterance, text (c) semantic per se

Utterance, text, situational context (d) implicatiortal: sense per se

A s soon as th e sp ea ke r is given th e floor, th e interp reter activates all the

levels i.e. syntax, sem antics an d p rag m atics s o th a t th e y c a n a c t coop era tive ly

to decip her th e speaker’s incom ing m essage. D raw ing o n Ch erno v, p robability

prediction begins in tw o directions: top-dow n and bottom -to-top. If the speaker

Is know n to the interpreter, he can m ake the use o f situational factors,

background know ledge and exp erien ces to understand the speaker's line o f

reasoning. T h is process in Ch erno v's w o rd s is called " t o p - d o w n " . V isual

Information such as the facial expression s o f the spea ke r and a v ie w o f the

conference hall are also considered part o f th is process. But if the spea ke r is
22 Transition Studies, Vol. 2, No. 7 & 8, Autumn 2004 & Winter 2005

unknown to the interpreter and the situation in g e n e ra l is unfam iliar, probability

prediction begins In the b o t to m - t o - to p d ire c tio n a n d s /h e sh ou ld rely on ^

sounds, words and phrases or first fe w sen te n ce s (th e m a tic p a rt) to understand

the upcoming linguistic streams.

The two processes interact and lack o f inform ation on one level can be
compensated for by checking against information on the other. Top>down
processing, in particular, plays an important role fo r simultaneous interpreters
working with asymmetrical languages since they cannot w ait for the end of an
incoming sentence before beginning to interpret.

The SI literature shows that the fo rm e r s tra te g y is m o re reliable and

leads to fewer errors or omissions. (N iska, 1999) T o be g o o d a t using the top-

down strategy, the interpreter should activ e ly re fe r to or rely on his

extralinguistic knowledge during the co urse o f th e liste n in g p hase. Highlighting

the inherent nature o f top-down stra te g y w h ile listen in g , David Nunan

(1999:201) writes quite precisely that

A listener actively constructs (or, m o re a cc u ra te ly reconstructs) the


original meaning o f the speaker usin g in c o m in g s o u n d s a s clues, in this
reconstruction process, the listen er u s e s p rio r k n o w le d g e o f the context
and situation within which the listen ing ta k e s p la ce to make sense cf
what he or she hears. C o n te « o f situ a tio n in clu d e s such things as
knowledge of the topic at hand, th e sp e a k e r o r speakers and their
relationship to the situation as w e ll as to e a c h o th e r a n d prior events.

As far as simultaneous interpreting is c o n ce rn e d th e practisearchers have

referred to different sources o f inform ation th a t s e rv e a s cues for the

interpreter to make guesses about the d ow nstream e le m e n ts o r predicates. As

regards these cues, the researchers have un an im ou sly c lassified them into two

types: linguistic and extralinguistic. T h e fo rm e r is m ad e w ith reference to the

purely linguistic (syntactic and sem antic) k n o w le d g e p ro v id e d b y the source

language sentence. According to W ilss (1978), lin guistic an ticipa tion is triggered
Anticipation Strategy and Simultaneous Interpreting 23

by certain linguistic units (e.g. words or word combinations), which serve as

cues. This Is found in the following example:

( l)
Bei allem, was sonst umstritten ist, meine ich Sp.:

In all that Int.:


mich also insoweit in sachlicher Übereinstimmung Sp.:
has been contested otherwise, 1 think that int.:
mit der Inhaltlichen Auffassung der Fraktion, dieses Sp.:
I am along the lines of the Int.:
Hohen Hauses zu befinden. Sp.:
General opinion of the parliamentary groups in the Bundestag. Int.:

(W ilss 1978: 349-350)

W here, according to Besien, th e w ord group m e in ic h m ich serves as a

cu e for w hat follow s no t In the least because the final verb (zu befinden) Is a

pallid verb w hich adds no further information to the sentence.

A sim ilar one is exem plified in (2) taken from Sha'bani (2001: 71):

(2)
ha r s c e n a rio • d a ra-v e c h a h a r bakhsh*e asasi Sp.:
each scenario having four parts fundamental Trartsc:

hast + yeki m a sa le -y e b ee ste la h - v a z ia t-e Sp.:


is one issue-of so-called state Transc.:
( lm a in £ o m P O N £ N IS [l-_ o n s.o rQ is Int.:

In exam ple (2) th e interpreter takes th e advantage o f his linguistic

know ledge o f th e collocation and form ulaic expressions and makes a prognosis

o f th e follow ing pallid verb w h o se first elem ent appeared in advance ju st after

th e subject.

By having a g lance a t an y language, one can find such pallid verbs that

are com prised o f d ifferen t elem ents. Usually one o f th e elem ents carries the
24 T aasM oa S tudie s, Vol. 2, No. 7 & 8, Autum n 2004 & W inter 200$

semantic load of the whole phrase and the others a re sem antically neutral in

linear representation of the sentence, the m eaningful elem en t ¡s placed ea1ier

which can trigger the appearance o f the second or th ird elements. A c o m p ^

listener can easily guess the neutral elem ents afte r receiving the meaning^

part. These elements of the verb are usually separated from each
stretches of intervening text Some o f such pallid verbs In Persian are n.
c »"fin
below:
I- beyangar-e.....boodan (to b e expre ssive o f)
2• stom ebe.......boodan (to in clu d e )
j - havi-ye .........boodan (to in c lu d e )

4- dara-ye......... boodan (to h a ve )


5- naghsh........ bazikardan (to p la y a ro le )
6- masooBat.......beohdeh g e re fta n (to u n d e rta k e a n o b lig a tio n )
neshane-ye..,..boodan (to be a s y m b o l o f)
$- gooya-ye ...... boodan (to be in d ic a tiv e o f)
9- ta'sir............. dashtan (to h a ve a n e ffe c t)
10-maz . dashtan (to ne ed )
n-khastar-e....... shodan (to re q u e st)
12-dshegh-e ...... boodan (to b e fo n d o f)

Also in the following exam ple th e m e an in g o f th e p allid verb is virtual

given in advance.

(3) M5B (may I just indicate that President. ..)


Sp.: Froher Leiter d er - S tr a fre c h ts a b te iu n g - in M ü n c h e n w ar bevor«
1983
Transc.: Earlier (leader of the penal law departm ent - in - Munich was
{before he 1983-
int: th a t . # - P resident O derskey - used to head,
Sp.: Pras. Des - B ayrischen (...) g e n ic h ts u n d d a m n 1988 P r .-d B
bundesh wurde
Transc: Pres. Of the Bavarian [...] court and then 1988 pres. €f
the Fededrat court became
Int.: and , he @ w as al$Q th e p re s id e n t o f the 8am an - W
Court

( A d a p te d fro m Setton, 1999:54)


Anticipation Strategy and Simultaneous Interpreting 25

In the above example, an actual predicate before the input appears, and
the first verb 'head' is cued by the meaning of 'Lelter' (leader).

W hile the Interpreter Is deciding on anticipating the SI input verb, s/he


should also think of its time or aspect. The interpreter can sometimes refer to
or get help from the textual clues, though not always present, embedded in the

sentence Just before the input verb. They can trigger the time of the utterance.
The examples of such clues are adverbs, adjectives, dates and word meanings.

(4)
Jin itia n ...yao tam tan (today,.A want to talk about...) Tsl-2:
Cong aanaoang de jiw e l jia n g zh e de yanjiang...
from (the) \just-now ...[ Speaker's speeches...

(Setton, 1999: 53)

In th e above example, th e tim e is marked by the words ganggang (just

now) and jin tla n (today).

T he o ther type o f anticipation is usually extrallngulstlcally oriented and

known to be backed by a w ide range o f sources o f information namely; world,

subject and situational knowledge, knowledge o f the speakers and social

conventions, etc. This anticipation is exemplified in (5), where the rapid

inference allow ing anticipation o f both th e final verb and its negation reflects

pragmatic inference as well as knowledge of social conventions as part of the

interpreter's extralinguistic knowledge:

(5) Speaker's previous sentence: Professor Odersky is going to speak on this theme.

als P rä s id e n te n >d e s 3 4 .7 deutschen 'B u n d e s Sp.:


as President ( AccJ - o f the G em an Federal Transc.;
... going to address precisely that/TH E M E 35.4 - In t.:
G e rich tsh o fs 36.3 - b rauch e ic h H errn - Professo r Sp.:
Justice Court - need I Mr [Acc] - Professor Transc.:
fte 3 6 .9 - hardlv 3 7 .7 needsanv Int.:
26 Translation Studie s, Vol. 2 , N o. 7 & 8, Autumn 2004 & Winter 2005

M trtfv « I — htor —MICKT VORiUStflllCfl 40.1 Sp.;


Otietsky h e re -D B U im lu S S Tr*^
^ ^ > » ¿ » 1 M 4 ¡think ah for 40 as you know he's. the. President o f the
FederalHigh Court Int:

(Adapted from Setton, 1998:

Drawing on Setton (1998:180) when the speaker produces 'hardly n e ^

any introduction', the interpreter knows the Chairm an is about to introtfu*


Prof. Defers*)', but in terms o f semantic input for this particular utterance ^

has oniy the Accusative phrase 'a /s P rasi'denten - d e s de utsch en ‘Bunded. u*

marled fronting of the Accusative phrase signals th e availability of a context^

effect The interpreter knows Odersky who occupies an exalted position that is

quite known to all. Therefore, the deduction that th e m oderator does not need

to introduce him is made and formulated in less than 2 seconds.

Besien (1999: 15) gives the following exam ple fo r this type of

anticipation:

(6)
Namens meiner Fraktion darf ich den beiden Herren Speaker:
On behalf of my political Interprete:
Berichterstattern für die Arbeit, die sie aufgewendet Speaker
group I should like to thank the two Interpreter
haben, sehr herzlich danken. Speaker;
rapporteurs very cordially for their work. Interpreter

Where the segment ‘rum ens m e in e r F ra n k tio n d a rf ic h ... danken' is i

standard phrase, frequently used as a debate-opening g a m b it which acconing

to Besien Is part of the interpreter's conference d iscou rse knowledge.

Some authors have also mentioned an othe r ty p e o f strategy for the

simultaneous interpreter when engulfed w ith th e sy n ta ctic constraints of the

output language. They refer to it as 'approxim ation' w h ich is not the exact M

an almost correct or near translation of th e source constituent. According to


Anticipation Strategy and Simultaneous Interpreting 27

Besien (1999), sometimes several approximations are given for the same
source constituent. The point to bear in mind is that the occurrence of this kind
of translation Is not due to the shortage in the interpreter's language
competency or translation ability but mostly to his short horizon in terms of
segments available for processing, pressure o f time and the need to divide
attention. (Gile, 1999)

In addition to the aforementioned way of looking at anticipation strategy


in terms o f linguistic or extralinguistic, SI literature also witnessed an alternative
approach. Lederer (1981), for instance, makes a fine distinction between 'pure,
observable' anticipation as opposed to 'freewheeling' anticipation. He holds that
the former is made when the interpreter produces the output constituent before
receiving the corresponding input constituent, but the latter happens when the
interpreter produces a segment in the output language almost at the same time
or a little later than the corresponding segment in the output language is

uttered by the speaker, but it is so soon that undoubtedly the interpreter had
guessed it before hearing the original. The interpreter decides on his

anticipation o f the speaker's unfinished utterance, based on what he has at his

disposal, and then listens to the speaker merely as a control and the translator
occurs in th e blink o f an eye. AN the examples coded so far can be categorized

into the class of pure or observable anticipation but the following extract from
Lederer (1981:252) is of the freewheeling type.

Speaker: (7 )
... Des­
halb
nen Ihrer eigenen Prototypen, mit der Ausstattung Quinet,
Z u r V erfügung stellen könnte (could place a t our
disposal)
so dass am 16. Januar der Presse
auch die Variante Quinet gezeigt warden könn­
te. Oie Schwei2e*
28 Translation S tudie s, Vol. 2, No. 7 & 8, Autumn 2004 8c Winter 2005

in te rp re t
11serait partfculiè*
Remit souhaitaWe que la SNCf
puisse mettre à notre disposition un d e se s
prototypes ayant un aménagement Quinet,
de manière è ce que le 16 janvier on puisse également m ontrer
èia presse
la variante Quinet.

D iscussion a n d c o n c lu s io n

Throughout the paper, the main ambition pursued was, first and foremost, to

present a full sketch of anticipation proper and treat at multivariate levels tg

show its prevalent occurrence and wide usage in the com plex cognitive task of

interpreting in general and simultaneous Interpreting In particular, it was

explained that language communicators benefit from a n anticipatory abffity

while processing an aural message to arrive at m eaning. T h is was also true of

simultaneous interpreters in SI task as the main core o f the present study. It

was noted that a successful interpreter is the o n e w h o is communicatif

competent in his SI in terms o f syntactic, sem antic an d pragmatic knowledge

Such a complex knowledge can be taken as a reliable so urce o f information to

hypothesize on the probable linguistic am biguities emanating from the

insufficient amount of input at the time o f interpreting w here an immediate

rendition is mandatory. As explained, w hat m akes this immediate response

more risky is ttie asymmetrical language com bination which forces the

interpreter to delay his output because o f th e typological straitjacket of the

surface structure of output language. Prim arily, tw o typ e s o f knowledge

supported the interpreters in their prediction an d an ticipation that are classified

into linguistic and extralinguistic. The form er is subcategorized into syntactic

and semantic knowledge, that is, the know ledge o f S O V word order and

co0oc3tiOfl, idioms or formulaic expressions. T h e latter is also subcategori»*

into the knowledge of speed) acts, speech events, co ntext or situation and
Anticipation Strategy and Simultaneous Interpreting 29

schematic knowledge which is the product of the human's background


experiences, social group, culture, cognitive scripts and frame knowledge.

These packets of knowledge provide a basis for the Interpreter to predict


the speaker's downstream strings. A plausible schematic projection is the result
of the presence of precise assumptions plus their on-line activation and lack of
either one can lead to a wrong anticipation or error. The interpretation
examples taken from the live corpus showed that the interpreters mostly
resorted to top-down strategies and decoded the incoming message at a
macro-level. To uncover the meanings, they took the advantage of the helpful
cues ranging from linguistic to extralinguistic knowledge that triggered the
occurrence o f next elements. As a suggestion for prospective interpreters,
preparation and acquaintance with the subject matter to be discussed or even
meeting the speakers before the conference could make the interpreter ready
and put him/her at the right path so that s/he can make correct, Intelligent
guess and, in so doing, foster his/her anticipatory capacity to generate

predictive inferences.

The availability of potential knowledge is also not sufficient for

comprehension; potential knowledge must be activated in order to facilitate


Interpreter's ability to understand.

T o arrive at an immediate decoding and interpretation o f the on-going

message delivered by the speaker, the Interpreter would be better off


dismissing the neutral, redundant elements carrying no information such as
unmeaningful parts o f idiomatic expressions and o f pallid verbs or unpredictable
elements o f collocations.

In short, the knowledge of components o f context and its specific


variables should be fed first. These contextual factors Involved in simultaneous
interpreting situation are ’who is the speaker?*, *what is he talking about?', 'In
what connection is he speaking?', 'whom is he addressing?', 'where is he
30 T ra n sa tte * S tudie s, Vol. 2 , N o. 7 & 8, Autum n 2 0 0 4 & W in te r 200S

speaking?', 'when is he speaking?', *what is h e a im in g a t? ' and *why |S ^


speaking?' An interpreter hoping to g u arantee a s u c c e ssfu l rendition sh<m
have the answers to above questions before th e c o n fe re n c e starts. Otfietv^
he would dig himself into a hole.

W orks C ite d
AMoson, J. C. (1980). Native and Nonnative Speaker Perform ance on Ctoze Te^
language Learning. 3 0 ( 1) S9-76.

Anderson A. and Pearson, D. P. (1988). 'A Schem a-theoretic V ie w o f Basic Process®!,


Reading Comprehension', in Patricia L Carrell, et ai (1988). Int&dcfyt
Approaches to Second Language R eading, pp. 3 7 -5 5 . N.Y.:
University Press.

Anderson, A. & Lynch, T. (1988). liste n in g . Oxford: O xford University Press, pp. ^
259.

Sesìen F. V. (1999). Anticipation in Simultaneous interpretation. Meta, XLIV, 2. JS&.


259.

Bransfbrd, ID., & Johnson, M.K. (1972). Contextual Prereq u isite s for Understand^
Some Investigations of Comprehension and R ecall. J o u rn a l o f Verbal Leata^
and Verba! Behavior J i, 717-726.

Brown, 6. (1990). listening to Spoken English. Second Edition. London: Longman.

Chemov, G.V. 1994. Message Redundancy and M e ssag e A n ticip a tio n in Simultaneo»
Interpreting. In: S. Lambert & B. M oser-M ercer (e d s.) B ridging Me %
Empirical Research in Sim ultaneous in te rp re ta tio n . A m sterd am / Philddelpfc
John Benjamins.

Fromkin, V. and R. Rodman (1983). An In tro d u c tio n to Language. Newyork: Hot


Rinehart, and Wingston.

Gwmn, M. (1990). Psycholinguistics. Cambridge, C a m b rid g e U n ive rsity Press.

Gilè, D. (1992) Predictable Sentence Ending in Ja p a n e se and Conferva


Interpretation. Special Issue n o .l, The In te rp re te r's N e w sle tte r, 12-23. Trie#
università DelgJ studi Di Trieste.
Anticipation Strategy and Simultaneous Interpreting 31

Gile, D. (1999), Testing the Effort Model’s Tightrope Hypothesis In Simultaneous


Interpreting: A Contribution, H e rn s, Journal of Linguistics, no. 23, 156-172.

Lederer, M . (1981). La in d u c tio n sim ultanee. Experience e t theorie. Paris Mlnard.

U nderholm , T . <2002). Predictive Inference Generation as a Function of Working


M em ory Capacity and Causal T ext Constraints. In D iscourse Processes, 34(3),
2 5 9 -2 8 0 . Lawrence Erlaum Associates, Inc.

M ahm oudzadeh, K. (1994). Bilingualism, Interpreting and Mental Processes.


P roceeding o f th e Z * conferen ce on Translation, University of Tabriz. Pp, 103-
116.

M oilanazar, H. (2001). P rin ciple s a n d M e tho do log y o f Translation. SAM T Publishing.

Niska, H. (1999). Text-linguistic Models for the Study o f Simultaneous Interpreting.


Stockholm University. Internet File: WWW.linQ.oro/textlinQuist.hfaDi/

Nunan, D. (1999), S econd language teaching a n d learning. Hefnle & Heinie Publishers.

O iler, J . (1979). Language T ests in School. London: Longman.

Richards, J. C. (1983). Listening Comprehension: Approach, Design, Procedure. TESOL


Q u a rte rly 17:219-39.

R ichards ) . C. et al. (1992). Longm an D ictio n a ry o f Language Teaching and A pplied


L ing uistics. Longman.

Schank, R., Abelson, R. (1977). S cripts, Ptans, Goals a n d U nderstanding: An in q u iry


in to H um an K now ledge S tructures, New Jersey: LEA Publishers.

Setton, R. (1999). S im ultaneous In te rp re ta tio n : A C ognitive-pragm atic Analysis. John


Benjam in Publishing Company, Amesterdam/ Philadelphia.

Setton, R. (1998). Meaning assembly in simultaneous interpretation. In In te rp re tin g


Vol. 3(2), pp. 163-199. John Benjamins Pubflshlng Company.

Sha'banf, K. (2001). Investigating some Language-specific Problems in Farsi-To-Eng/ish


Sim ultaneous Interpreting. M A thesis. Tabriz University Press.

Sha'bani, K. (2004), Persian Verb-last Sentences In Simultaneous Interpreting.


T ranslation S tvdfesV of. 2, No.5, 7-32.
32 Transla tio n S tudie s, V ol. 2, N o. 7 & 6 , A utum n 2004 & Winter 2005

Tahowian, J. (1995), Linguistic and Extralingulstic Aspects of


Interpretation, In Motarjem, vol. 4, pp. 36-45, no.16.

van den Brock, P. (1990). The Causal Inference Maker: towards a Process Model *
Inference Generation in Text Comprehension. In D. A. Baiota, G. 8. Flo^
d'Arcais, a K. Rayner (Eds.), Comprehension Processes in reading (pp.^
445). Hillsdale, NJ: Lawrence Associates, Inc.

Wife J. (1981), Teaching English through English. London: Longman.

Wfas, W. (1978). Syntactic Anticipation in German-English Simultane«*


Intefpretatton. In David Gerver and Wallace H. Slnaiko (eds). Proceeding f t
NATO Symposium on Language Interpretation and Communication. Venice.
1977, 115-43. Newyork: Plenum Press.

Yule, G, (1996). Pragmatics. Oxford University Press.

Zotirevandl, Z. (1994), The Impact of Schematic Anticipation on Translation.


Proceeding o f the Second Conference on Translation, University of Tabriz. Pp.
145-158.

Speaker:

aden werden. Ferner beabsichtigen wir 58 1

dem Veiwaltungsrat gelegentlich der Sitzung 59

am 28. März in Basel je einen 60

Prototyp mit Jeder der beiden Varian­ 61

ten der Inneneinrichtung vorzuführen. 62

Interpreter:

et nous avons également l'in­ 59

tention le 28 mars à Bâle 60

de présenter les proto­ 61

types heu avec les différents types d'aménagement 62

intérieur, et en outre de heu des 63


On Machine Translation for Persian1

__________________________ Mehran A. Taghvaipour“

Abstract
This paper is an a ttem pt to c la rify th e aim s o f M T and high lig ht some o f the
reasons w hy Persian M T Is now a m atter o f necessity fo r Iranians in term s o f its
scientific, philosophical, social, and com m ercial benefits. A fte r giving a short
h isto ry o f MT, I w ill Illu stra te th a t translation qu ality Is n o t a ll th a t m atters In
th e 21st century. I w ill show th a t there are some other factors Involved th a t
m ake M T an attractio n fo r translators and non-translators the same. I w ill
b rie fly review a fe w types o f M T system s. From th e rule-based approaches,
transform er and lin g u istic knowledge (LK) architectures, and from the em pirical
approaches, exam ple-based and sta tistica l architectures are reviewed to give
th e reader an Idea about how M T engines work. Linguistic and/or com putational
problem s o f each approach are discussed in short. I t is argued th a t M T systems
a re b e tte r to use a com bination o f various architectures ra th e r than sticking to
a single approach. A s an exam ple o f such architectures, Bond's (2002) proposal
fo r a M ulti-Pass M achine Translation (MPMT) is reviewed. W herever possible,
te ch nicality is avoided a n d references are m ade to th e Persian language so th a t
th e m a te ria l is m ore a ttra ctive and tangible fo r non-M T practitioners.

Key Words
Persian Machine Translation, Computational Linguistics, Translation, Computer
Sciences, Artificial Intelligence

I This paper was received on 11/12/2004 and approved on 14/02/2005.


II PhD student in Computational Linguistics at the Department o f Language and
Linguistics, University of Essex, UK; e-mail: mtaQhvaioouf@hotmail.com

You might also like