The unbearable lightness of English words


Andrew Chesterman

                     [2007. In R. Jääskeläinen et al. (eds), Text, Processes, and Corpora: Research Inspired by Sonja Tirkkonen-Condit. Joensuu: Joensuun yliopistopaino, 231-241.]




The paper postulates a “rhetorical salience threshold” which may have different heights in different languages. This threshold marks the point at which a given item of information or component of meaning is judged to be salient enough to be worth expressing. In translation, if source and target languages have different salience thresholds, rhetorical adjustments may need to be made which have the effect of “toning down” or “toning up” the salience. If there is no such rhetorical compensation, the translation may sound either “too pompous” or “a bit pathetic”. Evidence is offered which suggests that Finnish and English have different salience thresholds: the English threshold seems lower than the Finnish one.



1. The initial question


An English TV ad for a shampoo says that if you use it your hair will be Fuller, thicker, fresher, cleaner — better than ever. When this same ad appears on Finnish TV something happens to the adjectives: three of them lose their comparative status, thus: Raikkaat, puhtaat, terveet hiukset — kauniimmat kuin koskaan (‘Fresh, clean, healthy hair — more beautiful than ever’).

A Finnish text in a translation exam mentions a research fund that has succeeded in accumulating a sizeable sum of money from scratch lyhyessä ajassa (literally, ‘in a short time’). Of 25 examinees, 14 indeed do translate the passage as something like “... the contribution of the fund is significant in that the University has been able to raise the sum from zero to the present level in a short time”; the examiner feels there is something stylistically odd about the end of the sentence. Eleven examinees feel the need to add something to the phrase: in such a short time or in a very short time; the published translation also adds very.



2. Thresholds of silence and salience


Research on silence has looked at different forms and functions of silence, and examined the different roles which silence can play in different cultures (see e.g. Tannen and Saville-Troike 1985, Laaksovaara and Farell 1992, Kukkonen 1993, Jaworski 1993). One well-established result is that Finns’ tolerance of silence is notably higher than that of, say, Americans. In other words, Finns tend to have a higher silence threshold than, for instance, speakers whose native language is English. This threshold is the point above which speakers feel it necessary to say something.

                  The approach taken by much of this research is semiotic: silence is seen as a sign like any other sign, with its own range of signifiés. Silence is not “just” silence, the absence of meaning; silence means something. And this “something” is determined by the whole cultural semiosphere in which a silence exists, for different semiospheres will create different contexts for silence, different frames of reference within which silence can have a meaning. The Finnish semiosphere is particularly interesting in this respect. Tarasti (1988) has argued that it is less profuse, more sparse than the semiosphere of some other cultures, especially when compared to the Anglo-American one. Some may see this as a weakness, a mark of semiotic poverty. But this is not necessarily so: in a sparse semiosphere, individual signs can be stronger, they can carry more weight.

                  Parallel to the culture-bound silence threshold I suggest there is also a culture-bound rhetorical salience threshold[1], defined as follows:


                  The rhetorical salience threshold marks the point above which something is felt to be significant enough to be worth saying.


Note the difference from the silence threshold: whereas that marks the lower edge of the felt need to say anything at all, the salience threshold marks the point at which a given signifié is felt to be worth uttering. The higher the silence threshold, the smaller the need to speak; the higher the salience threshold, the more meaningful something must be in order to become rhetorically salient enough to be uttered. Information below the salience threshold is deemed “insignificant”, not worth uttering. Admittedly, some information may remain unexpressed because it is deemed too important to utter: see e.g. Leppihalme 2000: 98-90.

Recall the examples given above. In both cases, we seem to see evidence of an adjustment of the salience threshold. A “closer” Finnish translation of the first one would give: Raikkaammat, puhtaammat, terveemmät hiukset. But for a native Finn, this perhaps sounds “too strong”, “over-written”. And a literal translation of the second example would seem, as I suggested above, somehow not strong enough, as if it were “written too quietly”. “Writing too quietly” may mean no more than “uninteresting subject matter” (for a given reader); but a text may also fall beneath the salience threshold for rhetorical reasons. Specifically, however, I am interested not in why texts might fall beneath this threshold, but in what writers do (in different languages: here, Finnish and English) to ensure that their texts do not fall below it, or that they do not rise too far above it.

                  I suggest that the salience threshold is lower in English than in Finnish. Evidence may be found in situations where a given signifié is shared (or assumed to be shared) between two texts, one Finnish and one English: we can then examine differences at the level of the signifiant. An obvious source of such data is translations, and another is newspaper reports in different languages dealing with the same incident.

                  The structure of my argument is as follows. I assume (a) that all language use must rise above the salience threshold, and (b) that languages may vary in the degree of salience that must be manifest in a signifiant in order for it to be worth uttering. I then argue that some signifiants in English may seem too “light” (i.e. not salient enough) to justify expression above the salience threshold unless they are supplemented in some way that compensates for this “lightness”. In Finnish, the use of signifiants for the same signifiés does not seem to require such supplementary compensation. I infer that (in these cases) Finnish signifiants are therefore intrinsically “heavier”, more salient.

                  The converse argument is also possible: i.e. that if certain Finnish signifiants are supplemented in the same way as corresponding English ones, the result will be a Finnish text that is felt to be “over-written” or cluttered, sensationalist or the like. This opens up possibilities of empirical reader-response testing that I have not yet explored.

                  But first, some caveats. Considerations of word length and word frequency are excluded here, although one might certainly argue that low-frequency items are less predictable, hence more informative, and hence given “supplementary weight” in some sense. And I also exclude differences which are solely determined by grammatical constraints, where the writer has no other stylistic choice available. In the present context I also omit discussion of the way the overall hypothesis might relate to different politeness conventions in the two cultures.

                  Furthermore, differences of situation and context must also be eliminated: it is obvious that in some contexts X might be worth saying and in other contexts not. Hence the usefulness of newspaper texts, from which I shall present some evidence: we can assume a large degree of similarity between newspaper-reading situations in Finnish and English, and the newspapers selected can be approximately matched on the quality-popular scale.



3. Some preliminary evidence


Some evidence for my general argument that Finnish words are “heavier” than English ones comes from a comparison of newspaper texts. When I first started investigating the the salience threshold some years ago, I took one issue of the national daily Helsingin Sanomat, of December 8, 1992, and the issue of The Independentof the same day. Both these newspapers are quality papers; neither is linked to any particular political party.

                  I then selected pairs of articles from the international news pages that dealt with the same events, each article being by a different writer. These were: rioting between Muslims and Hindus in India, the clash in the Russian Parliament between the supporters and opponents of Yeltsin, corruption scandals in Japan, and the final match of the Davis Cup (tennis).

                  I looked for specific signifiés (i.e. sememes) that occurred in both newspapers. The focus of analysis is on the differences in the ways these signifiés were expressed. These differences are assumed to indicate various ways of compensating for possible differences of salience. Let us look at some data first.


The article on unrest in India dealt with the destruction of a Muslim mosque. I will take one aspect of the journalistic treatment of this event in the two newspapers, by way of illustration. Since I am interested in qualitative differences rather than quantitative ones, and the data are very limited, frequency figures are not given.

                  Both articles naturally make repeated reference to the process in which the mosque was the object, in the semantic role of Patient. The English text uses the following verbs or nominalizations to denote this process: demolish, destruction, overrun, attack, demolition, destroy, smash with sledgehammers, storm. Compare this list with the equivalent expressions used in the Finnish article: tuhoaminen (‘destroying, destruction’), tuho (‘destruction’), purkaminen (‘taking to pieces, dismantling, demolishing’), hyökätä (‘attack’). No other verbs or nominalizations are used to express the process in question, in either text.

                  It seems that the Finnish items are, on the whole, somewhat less emotional, less marked. In particular, the neutral sense of ‘taking to pieces’ (which actually occurs twice in the Finnish) is noticeably absent from the English article, which seems to make more of a point of selecting stronger terms. The English also gives more physical detail (sledgehammers).

                  The headlines of the two articles are also revealing: the English was: Sectarian rioting kills 200 in India. This focuses on a transitive verb, in the present tense, and selects rioting as the agent. The Finnish had: Satoja ihmisiä kuollut Intian levottomuuksissa (‘Hundreds of people dead in riots in India’). This has an intransitive verb and refers to the riots literally as “incidents of unrest”). One might say that riot is a hyponym of “unrest”, a specific type of it, marked e.g. for high conflictual content.


The pair of articles on Japanese corruption scandals exhibited a similar tendency. Consider, for instance, the terms used to describe the group of criminals allegedly involved. The English refers to these as: gangsters, gangster syndicates, a yakuza gangster syndicate. It also uses the term gangster row in the headline.

                  In referring to these same groups, the Finnish text uses järjestäytynyt rikollisuus (‘organized crime’), jakuzat and rikollisuusympyrät (‘criminal circles’). It does not use the term syndikaatti, a possible but low-frequency item, and prefers a more neutral mode of expression; it also avoids any close equivalent for gangster. The Finnish headline does mention rötösjutut, which can be literally translated as ‘criminal matters’ but has cultural-specific connotations of corruption; but note that whereas this Finnish term refers only to events the equivalent in the English headline includes a premodification referring to human Agents: gangster row.


The articles on the Russian Congress debates offer rich material for analysis along the same lines. I will pick out two examples.

                  The first is the modes of reference to the various groups of Congress representatives involved. Here are the terms used in the English text: hardline Communists, belligerant nationalists, conservatives, rival camps, stodgy veterans, rival forces, centrist deputies, opponents. Note how many are preceded by emotive attributes.

                  The Finnish text (slightly shorter) used a more limited set of terms to denote the same participants: vanhoilliset (‘conservatives’), enemmistö (‘the majority’), vanhoillinen enemmistö, (‘the conservative majority’) erittäin suuri enemmistö (‘an extremely large majority’). It is interesting that the English text at no point uses “the majority” to refer to the representatives themselves: the term is only used in its technical sense, a two-thirds majority.

                  The second example has to do with the expressions used to describe the emotional atmosphere of these debates. The English article had: an all-or-nothing showdown, dramatic sabre-rattling, unruly and unpredictable Congress, head-on confrontation, within a hair’s breadth of a humiliating defeat, loathed by many stodgy veterans, political deadlock, political paralysis, back-room haggling, political pyrotechnics. The Finnish text had: tunteita kuumentava kysymys (‘a question which heated feelings’), sekoileva kongressi (‘a congress that was making a mess of things’), julkaisi yleisen tuomitsevan lausuman (‘made a general condemnatory statement’), vastustivat kynsin hampain (‘resisted tooth and nail’), ehkä varmistaakseen epäjohdonmukaisuutensa (‘perhaps in order to ensure their inconsistency’), punainen vaate vanhoillisille (‘a red rag to the conservatives’), tulenhehkuvan riidan syy (‘the cause of a fiery conflict’). Some of these Finnish expressions are strongly emotive, it is true, but the general effect overall is clearly less marked than the English.

                  The tone of both articles is set by their headlines. Whereas the English has Unruly Russian MPs sheath their sabres, Finnish settles for a matter-of-fact Jeltsinin vastainen rintama supistui kongressissa (‘The front opposing Yeltsin diminished in Congress’). Again, Finnish shows a preference here for a verb that is syntactically intransitive and a neutral tone.


My final text was a sporting one, and this turned out to be the most revealing of all. Not only did both newspapers treat a distinct series of events differently, but the Finnish even contained a translation of a quotation used verbatim in the English. The articles are reports of the Davis Cup tennis final, in which the USA defeated Switzerland.

                  The headline of the English article stresses competition and a positively regal supremacy: Courier predicts a long reign. The Finnish report stresses co-operation, with the headline: Courier on joukkuepelaaja (‘Courier is a team player’). The first paragraph of each report places the event in the context of the US defeat the previous year. But notice the different ways in which this is done. The English text talks of the present victory eclipsing the painful memory of their unexpected defeat by France in Lyons last year. In other words, this focuses on a personal emotional memory, and on the previous defeat being unexpected. The Finnish simply has: USA:n tennismaajoukkue paikkasi vuoden takaisen tappionsa Davis cupin loppuottelussa (‘The US national tennis team patched up their last year’s defeat in the Davis Cup final match’). Here, in contrast, the focus is non-personal, non-emotive, and the previous defeat is not given any epithet.

                  Elsewhere in the texts, both writers make use of the image of setting a seal on something. The Finnish says that Courier sinetöi (‘set the seal on’) the championship by winning his match, but the English use of the image is intensified to Courier ... put the final seal on a successful if sometimes erratic year ... Here the seal had to be final.

                  The difference between the two styles also comes out in references to previous defeats suffered by Sampras, another American player in the team. The Finnish text simply states bluntly that last year Sampras hävisi molemmat kaksinpelinsä (‘Sampras lost both his singles matches’). Compare the English version: His Davis Cup debut in last year’s final was a personal nightmare as he lost both of his singles matches.

                  The Finnish continues: Tällä kertaa hän esiintyi nelinpelissä John McEnroen kanssa (‘This time he appeared in the doubles with John McEnroe’). But the English adds a figurative image: tandem with John McEnroe. The Finnish description of this match goes: Sveitsin pari ... voitti kaksi ensimmäistä erää ... mutta kolme viimeistä menivät kotijoukkueelle (‘The Swiss pair ... won the first two sets ... but the last three went to the home team’); note the simple verbs. Compare the English: [Sampras and McEnroe] fought back to win the doubles after trailing by two sets to love ... ; note the emphasis on verbs of conflict.

                  The American Courier is quoted at the end of the article as having the utmost respect for the Swiss players; the Finnish translates the same quote as Arvostan ... hyvin korkealle (‘I respect [them] very highly’). “Very highly” is certainly high, but surely less high than utmost: here again, the Finnish is content with a slightly lesser degree of persuasion – and even this seems perhaps too strong.



4. Salience adjustment


Examples such as these suggest various ways in which writers can adjust the level of rhetorical salience to the relevant cultural norms. In the preliminary list of syntactic, semantic and stylistic/pragmatic variables below, “shift up” means “do this in order to increase salience”, and “shift down” means “do this in order to decrease salience”.


• Positive vs comparative vs superlative

                                   Shift up: move towards the right of this scale

                                   Shift down: move leftwards.

                                   Example: the shampoo ad

• Presence or absence of adjective

                                   Shift up: add adjective

                                   Shift down: omit adjective

                                   Example: unexpected defeat vs tappio (‘defeat’)

• Presence or absence of intensifier

                                   Shift up: add intensifier

                                   Shift down: omit intensifier

                                   Example: lyhyessä ajassa —> in a very short time

• Transitive vs intransitive verb

                                   Shift up: choose strong semantic role, e.g. Agent

                                   Shift down: choose weaker semantic role, e.g. Dative

                                   Example: kills vs on kuollut (‘has died’)

• Presence or absence of emotive feature

                                   Shift up: add emotive feature

                                   Shift down: remove emotive feature

Example: stodgy veterans vs vanhoilliset (‘conservatives’)

• Figurative vs literal

                                   Shift up: choose figurative expression

                                   Shift down: omit figurative expression

                                   Example: sheath their sabres vs supistui (‘was reduced’)

• Hyponym vs superordinate term

                                   Shift up: choose lower-level, more specific, more informative term

                                   Shift down: choose superordinate term

                                   Example: smash (with sledgehammers) vs tuhota (‘destroy’)

• Trope preference

                                   Not examined in any detail yet, but salience might be increased by conflict metaphors; compare emotiveness.

                                   Example: fought back to win vs menivät (‘went’)    

• Use of register

                                   Shift up: choose a down-to-earth, colloquial register

                                   Shift down: choose a neutral register

                                   Example: stodgy vs. vanhoillinen ‘conservative’



5. Conclusion


The tentative hypothesis outlined above – the concept of the salience threshold, and its explanatory power in accounting for some translation shifts – may be linked to the results of some other research. Séguinot (1982), for instance, studied the kinds of editing changes made to journalistic texts by translators working from French to English. She found that, after changes made to improve readability, the second most frequent category was composed of changes that included the reduction of emotive and figurative language. This seems to suggest that the salience threshold is lower for English than for French, as the French originals needed to be “toned down” to meet the acceptability norms of English. Vehmas-Lehto (1989) notes the necessity of a similar toning down in translation from Russian to Finnish.

Jantunen (2001) studied synonymity and lexical simplification, using part of the Savonlinna Corpus of Translated Finnish (see Mauranen 2000 for details). He analysed a subcorpus consisting of 1.2 million tokens of original Finnish and 1.2 million tokens of Finnish translations from English. One of his results was that there is a difference in the use of boosters (words like erittäin ‘extremely’, hyvin ‘very’, todella ‘really’) in translated and non-translated texts, particularly in narrative prose. His table 1 shows that in translated Finnish, boosters occur about twice as frequently as in original Finnish. He does not speculate on possible reasons, and does not give examples, but I find his result a striking one. I suggest that what his data show is an effect of interference: perhaps the equivalent boosters existed already in the English source texts (not present in the corpus), and were simply transferred as such into Finnish, without any adjustment to the salience threshold.

All the evidence I have discussed is only circumstantial, at best. Some of the differences noted in the journalistic texts may be specific to particular journalistic conventions, and this is obviously something that should be checked against other material. Yet the notion of the salience threshold, and the hypothesis that in English this tends to be lower than in Finnish, both seem interesting enough to warrant further study.





Chesterman, A. 1997. Memes of Translation. Amsterdam: Benjamins.

Jantunen, J. H. 2001. Synonymity and Lexical Simplification in Translations: A Corpus-based Approach. Across Languages and Cultures 2: 1, 97-112.

Jaworski, A. 1993. The Power of Silence: Social and Pragmatic Perspectives. Newbury Park, CA.: Sage.

Kukkonen, P. 1993. Kielen silkki. Hiljaisuus ja rakkaus kielen ja                kirjallisuuden kuvastimessa. Helsinki: Yliopistopaino.

Laaksovaara, T.H. and Farell, G. 1992. Position of Silence in English and Finnish Culture. Erikoiskielet ja Käännösteoria  / VAKKI-symposium XII, 107-118.

Leppihalme, R. 2000. Kulttuurisidonnaisuus kaunokirjallisuuden kääntämisessä. In Paloposki, O. and Makkonen-Craig, H. (eds), Käännöskirjallisuus ja sen kritiikki. Helsinki: AKO. 89-105.

Mauranen, A. 2000. Strange Strings in Translated Language: A Study on Corpora”. In Olohan, M. (ed.) Intercultural Faultlines. Research Models in Translation Studies I. Textual and Cognitive Aspects. Manchester: St. Jerome Publishing. 119-141.

Séguinot, C. 1982. The Editing Function of Translation. Bulletin of the Canadian Association of Applied Linguistics 4: 1, 151-161.

Tannen, D. and Saville-Troike, M. (eds) 1985. Perspectives on Silence. Norwood, NJ: Ablex.

Tarasti, E. 1988. Suomi semiootikon silmin. Synteesi 1-2, 12-19.

Vehmas-Lehto, I. 1989. Quasi-correctness. A Critical Study of Finnish Translations of Russian Journalistic Texts. Helsinki: Neuvostoliittoinstituutti.


Back to Publications List

[1] In Chesterman 1997: 114 I called this notion the significance threshold, which was not a good term as it is easily confused with statistical significance.