Using translation to show how the perfect differs across languages

Many western European languages have a perfect tense, formed by combining an auxiliary verb (meaning ‘have’ or ‘be’) with a past participle. Different languages use this verb form in different ways. A recent paper used translations of a well-known French novel to explore those differences. The aim was to see which tense the translators used in translating French verbs in the present perfect tense. The paper is interesting for several reasons:

  • it explores how usage of perfect tenses varies across several languages;
  • it considers how translators translate the present perfect tense;
  • it uses data from translated literary fiction to map out differences between languages;
  • it uses a technique called multi-dimensional scaling to generate maps showing how usage varies between languages; and
  • it showcases a database and analysis tool called translation mining.

The paper A multilingual corpus study of the competition between past and perfect in narrative discourse, (by Martijn Van Der Klis, Bert Le Bruyn and Henriëtte De Swart) appeared in the Journal of Linguistics in 2022.

I discuss below:

  • tenses studied
  • method
  • frequency of the PERFECT
  • subset relationships
  • which uses in which languages?
  • limitation and follow up

Tenses studied

For tenses in individual languages, the paper uses the names used in traditional grammars (eg Passé Composé in French). To refer to categories of tenses across languages, the paper uses standardised labels in SMALL CAPITALS. (I can’t use SMALL CAPITALS in this post, so instead I use ALL CAPITALS). Table 1 shows:

  • the 3 cross-language categories of tense mentioned in this post.
  • the tense in each language, labelled with the names used for it in traditional grammars.
FrenchPassé ComposéPassé Simple
ItalianPassato ProssimoPassato Remoto
SpanishPretérito Perfecto CompuestoPretérito Indefinido
EnglishPresent PerfectSimple Past
Note: VTT = Voltooid Tegenwoordige Tijd; OVT = Onvoltooid Verleden Tijd
Table 1. Cross-language tense categories, and names of tenses in individual languages
Source: Source: van der Klis et al (2022) [extract]

French, Italian and Spanish (which are all members of the Romance family) distinguish a PERFECTIVE PAST (reporting foreground events) from an IMPERFECTIVE PAST (reporting background states). German, Dutch and English (Germanic languages) do not make that distinction and so their past tenses are regarded as instances of (SIMPLE) PAST.

The authors comment that Greek does distinguish between PERFECTIVE PAST and IMPERFECTIVE PAST, but the form of some verbs does not show this distinction (notably είμαι ‘to be’ and έχω ‘to have’. For convenience, they annotated those verb forms as Past and classed them as (SIMPLE) PAST.


In French, the Passé Simple is still used in literary and formal writing, but the Passé Composé has largely displaced it in conversation and informal writing. Similarly, the PERFECT forms have largely displaced (SIMPLE) PAST and IMPERFECTIVE PAST forms in southern German and northern Italy.  


The authors used a methodology they call Translation Mining. This involves creating a database containing a text and translations of that text. They tag each sentence (and each sentence in the translations) to indicate a property of the sentence. In this case, the tag indicated the grammatical tense of a verb in that sentence.

Source text

For this paper, the authors analysed the first 3 chapters of the French novel L’Étranger, by Albert Camus, together with published translations into 6 other languages: Italian, German, Dutch, (European) Spanish, English and (modern) Greek.

They picked this novel because, as is well known, the text is written largely in the present perfect tense (Passé Composé), instead of the past historic tense (Passé Simple) which is normally used in literary narratives of this kind.

Sentences analysed

The authors analysed all source text French sentences containing a verb in the present perfect tense (Passé Composé). They analysed the sentences in 3 stages:

  • first, they looked at how often each translation used a perfect tense to translate the French perfect tense;
  • second, they considered whether uses of the perfect tense in one of the 7 languages was a subset of the uses in the other languages.
  • third, they looked for cut-off points between uses of the perfect in one language and uses in the other languages.

Frequency of the PERFECT

The authors found 348 instances of the French PERFECT in the first 3 chapters of L’Étranger. The translators also used the PERFECT in almost all of those instances in Italian (338) and a little less often in German (326). They rarely used the PERFECT in Dutch (39), Spanish (16), English (11) and Greek (1).

When the PERFECT was not used, the main other forms used were:

  • the PERFECTIVE PAST in Italian (5 instances out of 10), Spanish (331 out of 332) and Greek (328 out of 347)
  • the (SIMPLE) PAST in the Germanic languages German (20 out of 22), Dutch (305 out of 309) and English (328 out of 337).

As those numbers show, for each language, the translator for that language used a single tense for most of the instances where they did not use a PERFECT form.

Subset relationships

The authors used a map-colouring technique to investigate the relationships between the translations. Figure 1 summarises the result. A blue dot depicts a sentence where the translator used that language’s counterpart of the PERFECT and a green dot depicts sentences where the translator used a different form. The area surrounding the dots is coloured in the same colour. Next to each coloured area is found the traditional label for the corresponding tense.

For French, all the dots are, of course, blue. For Italian, there are only a few green dots, too few to show up as a separate area.

For German, there are only a few more green dots, but enough to show up as a contiguous green area.   

The green area is larger for Dutch and becomes progressively larger for Spanish, then English, finally coming to occupy almost the whole map for Greek.

Figure 1. Temporal maps displaying PERFECT and PAST use per language.
Source: van der Klis et al (2022)

The authors interpret figure 1 as showing that the languages are in a subset relation, as illustrated in figure 2:

  • uses of the PERFECT in Greek are a subset of its uses in English
  • uses of the PERFECT in English are a subset of its uses in Spanish
  • uses of the PERFECT in Spanish are a subset of its uses in Dutch
  • uses of the PERFECT in Dutch are a subset of its uses in German
  • uses of the PERFECT in German are a subset of its uses in French and Italian. The authors say that they do not have enough data to distinguish French from Italian.
Figure 2. Subset relations in the distribution of the PERFECT.
Source: van der Klis et al (2022)

Which uses in which languages?

The authors then analysed the boundaries between each subset and the subset immediately within it. The aim was to identify the factors that explained why the translator for one language used the PERFECT but the translator for the other language did not. The following sub-sections discuss each boundary.

From French and Italian to German

In 25 cases, German translation uses the Präteritum (20 cases) or other verb forms (5 cases), but French uses the Passé Composé. Examining those cases, the authors conclude that German does not use a PERFECT form for stative verbs (expressing a state, not an action). On the other hand, French and Italian can use the PERFECT for stative verbs used in narrating a sequence of events.

Here is one of the 3 examples they cite:

(1a) J’ai voulu fumer une cigarette à la fenêtre. [French]
(1b) Ancora ho volute fumare una sigaretta all finestra. [Italian]
(1c) Ich wollte eine Zigarette am Fenster rauchen. [German]
(1d) I wanted to smoke a cigarette at the window. [English]

In each example in this post I reproduce the French source text, the English translation and the translations into the languages on either side of the subset boundary. The verb form under discussion is set in bold type.

From German to Dutch

The authors identified over 180 cases where the German translator used a PERFECT form, but the Dutch translator did not. Here is one example.

(2a) Il s’est assis derrière derriere son bureau, il a croisé ses petites jambes. [French]
(2b) Er hat sich hinter seinen Schreibtisch gesetz, hat seine kurzen Beine untereinander geschlagen. [German]
(2c) Hij nam achter zijn bureau plaats en legde zijn korte beentjes over elkaar. [Dutch]
(2d) He sat down behind his desk and crossed his short legs. [English]

The authors identified the common factor in these cases as being narration of, for example, a sequence of events. The authors conclude that French, Italian and German permit use of the PERFECT in narrative sequences, but Dutch (and Spanish, English and Greek) do not.

This conclusion is consistent with the observation that the PERFECT in French Italian and German is developing into a (PERFECTIVE) PAST.

From Dutch to Spanish

In 24 cases, the Dutch translator used a PERFECT form, but the Spanish translator used a PERFECTIVE PAST (Pretérito Indefinido).

The authors comment that Dutch prefers to report past events in the Voltooid Tegenwoordige Tijd, (except in narrating a sequence of events). But the Spanish uses the Pretérito Perfecto Compuesto only if the underlying event occurred in the ‘extended now’, typically in the last 24 hours.

Spanish bans past-time adverbials with the PERFECT, whereas Dutch (like French, Italian, and German) uses those adverbials freely with the PERFECT, as illustrated in examples (3a-3d), where the past-time adverbial is underlined.

(3a) Il a perdu son oncle, il y a quelque mois. [French]
(3b) Hij heft zijn oom een paar maanden geleden verloren. [Dutch]
(3c) Perdió a su tío hace algunos meses. [Spanish]
(3d) He lost his uncle a few months ago. [English]

The authors state that temporal connectives—such as ensuite ‘after that’, puis ‘then’, alors ‘then’, à partir de ce moment ‘from that point on’— are in Spanish just as incompatible with the PERFECT as past-time adverbials are.

From Spanish to English

In 6 cases, the Spanish translator uses a PERFECT form (Pretérito Perfecto Compuesto), but the English translator used a SIMPLE PAST (Simple Past). The authors discuss 4 of those 6 cases.

The authors comment that English doesn’t use the Present Perfect in reporting a past event that the writer assumes is known to the hearer (events that are ‘pragmatically presupposed’). The authors discuss one example where the context (an earlier part of the novel) tells the reader that the event has occurred.

(4a) ‘On l’a couverte, mais je dois dévisser la bière pour que vous puissiez la voir.’ [French]
(4b) ‘La hemos cubierto. Pero desatornilllaré el féretro para que pueda usted verla.’ [Spanish]
(4c) ‘We covered her up. But I was to unscrew the coffin to let you see her.’[English]

In this passage, the protagonist has gone to the mortuary where his mother’s coffin was resting. The statement quoted above is spoken by the undertaker, who knows that the protagonist has already seen that the lid had already been screwed onto the coffin. So, the English translation cannot use the Present Perfect. In contrast, the Spanish translation can use the Pretérito Perfecto Compuesto, because the coffin was closed within the last 24 hours.

The authors discuss another example where, they say, knowledge of typical real-world events tells the reader that the event must have occurred. For this reason, the English translation uses the Simple Past. In contrast, the Spanish translation uses the PERFECT because the event occurred within the last 24 hours.     

(5a) ‘On les a eus!’ [French]
(5b) Les hemos ganado [Spanish]
(5c) ‘We thrashed them’ [English]

Here, the protagonist hears shouts from people returning from a stadium. World knowledge tells us that a game always ends with a winner and a loser. The authors claim that the question under discussion is identifying the winner, not identifying the fact that there is a winner.

From English to Greek

The English translator used the Present Perfect in 9 cases where the Greek translator used the Aoristos.  They are all examples of 3 well-known uses of the present perfect tense in English. I will illustrate those uses with the following examples I have used before, not with the authors’ similar examples:

There is a 4th use of the present perfect in English—the perfect of persistent situation (or universal perfect):
Kati has lived in Budapest since last year [and still lives there]
The other languages considered in this paper use the perfect in such cases, instead of the perfect. As a result, the perfect of persistent situation does not appear in L’Étranger. See below for a discussion of some follow up work on this point.

The authors found only 2 instances where the Greek translation used either the Parakimenos or the PAST PERFECT. Both instances are negative existential phrases—phrases describing an event that would have had current relevance but did not, in fact, occur. Here is the example using the Parakimenos:
(6a) Depuis huit ans, ils n’ont pas changé leur itinéraire. [French]
(6b) In eight years, they haven’t changed their route. [English]
(6c) Edo ki ochto chronia then echoun allaxei diadrome. [Greek]

Because of the small sample size, the authors do not attempt an analysis of the Greek PERFECT.


Figure 3 summarises the subset relations and the constraints coming into play at each boundary. It uses information from a diagram in section 4.6 of the paper, but formatted like a diagram in van der Klis, Goldschmidt, Tellings, Le Bruyn and de Swart (2018).

It shows that the PERFECT can:

  • have existential uses in all 7 languages. Apparently, the PERFECT cannot have other uses in Greek (although the authors do not have enough data to attempt a detailed analysis of Greek).  
  • have resultative uses (and, presumably, experiential and hot news uses) in all 6 languages other than Greek.
  • be used even when it refers to an event already known to the reader (a pragmatic constraint that blocks its use in English) in 5 languages (French, Italian, German, Dutch and Spanish)
  • refer to bounded events in 4 languages (French, Italian, German and Dutch)
  • have narrative uses with non-stative verbs in French, Italian and German
  • have narrative uses with stative verbs in French and Italian
Figure 3. Differences in distribution of the HAVE PERFECT
Source: data from van der Klis et al (2022), format based on van der Klis et al (2018)

Limitation and follow up

The authors acknowledge that their data does not identify cases where one or more of the 6 target languages uses the PERFECT, but the French source text uses a different test. They say that an extended dataset used in a separate unpublished manuscript An unstable category in an overall stable tense-aspect grammar of the western European languages (Van der Klis, Le Bruyn, de Swart, 2021) confirms both the subset relation identified in this paper and the robustness of the insights provided in this paper.

A further follow up appears in a conference paper Perfect variations in dialogue: a parallel corpus approach, Jos Tellings, Martín Fuchs, Martijn van der Klis, Bert Le Bruyn, Henriëtte de Swart Proceedings of SALT 32 [Semantics and Linguistic Theory].
That paper analyses translations from chapters 1 and 17 of an English novel Harry Potter and the Philosopher’s Stone. It differs from the paper discussed in this post in the following respects:

  • the source language is English, not French. Thus, the sentences studied are ones in which English uses the PRESENT PERFECT, not ones in which French uses the PRESENT PERFECT.
  • the 5 target languages include Swedish (and, of course, French) but do not include Italian or Greek (nor, of course, English).
  • Among the sentences studied are ones in which the English Present Perfect refers to a persistent situation. This use of the PRESENT PERFECT also exists in Swedish, but does not exist in French, German, Dutch or Spanish.  
  • the text contains much more extensive dialogue. The paper analyses dialogue separately from narrative text. From this, the authors draw some inferences about the use of the PRESENT PERFECT in dialogue—though it must be remembered that the text doesn’t include real dialogue but dialogue created by the novelist for inclusion in the novel.

Another follow up appears in Perfect variations in Romance Henriëtte de Swart, Cristina Grisot, Bert Le Bruyn, Teresa M.Xiqués 2022. Special Issue of Isogloss. Open Journal of Romance Linguistics 8(5)/14, 1-31. DOI:
That paper finds that Romanian uses the PRESENT PERFECT as broadly as French and Italian do; Catalan uses it more narrowly than those 3 languages, but a little more broadly than Catalan; Catalan uses it slightly more broadly than Spanish; and Portuguese hardly uses it all all (even more narrowly than Spanish and English).


A multilingual corpus study of the competition between past and perfect in narrative discourse, Martijn Van Der Klis, Bert Le Bruyn, Henriëtte De Swart, Journal of Linguistics, Volume 58, Issue 2 , April 2022 DOI:, Open Access article, distributed under the Creative Commons Attribution-NonCommercial-ShareAlike licence (, permitting non-commercial re-use, distribution, and reproduction, if the same Creative Commons licence is included and the original work is properly cited.

Perfect variations in dialogue: a parallel corpus approach, Jos Tellings, Martín Fuchs, Martijn van der Klis, Bert Le Bruyn, Henriëtte de Swart Proceedings of SALT 32: 022–043, 2022

Perfect variations in Romance Henriëtte de Swart, Cristina Grisot, Bert Le Bruyn, Teresa M.Xiqués 2022. Special Issue of Isogloss. Open Journal of Romance Linguistics 8(5)/14.

An unstable category in an overall stable tense-aspect grammar of the western European languages Martijn Van der Klis, Bert Le Bruyn, Henriëtte de Swart, 2021, unpublished manuscript.


The paper is one of many outputs of a project on Time in Translation: The Semantics of the PERFECT

The Perfect, chapter 68 of the World Atlas of Linguistic Structures, discussed at

There is a good, concise summary of some of the work by this group of researchers in a a poster Harry Potter and the Rosetta Stone: uncovering differences in tense use through parallel corpora. van der Klis, M, Goldschmidt, A, Tellings, J, Le Bruyn, B, & de Swart, H (2018). Digital Humanities Community Event, Utrecht.   

Leave a comment

Your email address will not be published. Required fields are marked *

%d bloggers like this: