Description
Research on the PERFECT usually focuses on the English present perfect, which has a continuative reading with stative verbs (e.g. “Mary has lived in London for five years”, cf. Nishiyama and Koenig (2010)). However, most European languages do not share this continuative PERFECT and prefer a PRESENT in these cases (de Swart, 2016). On the other hand, while English rejects PERFECT tense (and resorts to the PAST) with past time adverbials (“Mary (*has) arrived yesterday”) and in narration (“...and then Mary (*has) arrived”), languages like French and German do allow PERFECTS in thesecontexts (Schaden, 2009).
A cross-linguistic account of the PERFECT seems necessitated. Parallel corpora can help: translation equivalents provide form variation across languages while meaning stays stable. We propose an application of Wälchli and Cysouw (2012), who use multidimensional scaling (MDS) to generate semantic maps from corpus data. We however operate at the level of grammar instead of the lexicon: we apply their methodology to differences in tense use between languages (van der Klis et al., 2017).
We annotated translations of Albert Camus’ L’Étranger (known for its narrative use of the passé composé) to investigate variation in PERFECT use. Starting from the French original, we annotated the verbal translations in English, German, Spanish, Italian, Greek and Dutch and attributed tenses to these verb phrases. After applying MDS, in line with the literature we find French and German on one end of the spectrum, and English and Spanish on the other. Our methodology allows to investigate where other languages reside on this continuum (Figure 1).
Figure 1: Results of applying MDS to tense variation in Albert Camus’ L’Étranger. Each point (347 in total) represents a tuple of tenses, corresponding to a text fragment (containing a passé composé). The colour labelling is according to Dutch tense attribution. Demarcation lines show where languages switch between PERFECT (left side of the line) and PAST (right). The boxes show our interpretation of
semantic differences between the demarcation lines (on the left: use of PERFECT for this phenomenon, right: using PAST).
We can also zoom in on fine-grained details: we e.g. find German (contrary to French and Italian) requires a PAST with (cognitive) stative verbs (like ‘want’, ‘think’, ‘know’, e.g. “J’ai voulu voir maman tout de suite” is translated with “Ich wollte sofort zu Mama”). And while Dutch does not allow a PERFECT in narration, it does in bounded situations (e.g. “Ik heb een beetje gedoezeld” is “I think I dozed off for a while”), separating Dutch from Spanish, English and Greek. If we broaden our view to all VPs in L’Étranger, we find the only PERFECTS introduced by other languages are continuative present perfects in English (e.g. “Il y a longtemps que vous êtes là” becomes “Have you been here long?”).
Stative verbs thus appear at the edge of the PERFECT in two ways: they are ingredients of both the PERFECT-PRESENT and the PERFECT-PAST competition. A preliminary collostructional analysis further suggests stative verbs are less likely to appear in a PERFECT per se.
| Period | 31 Aug 2018 |
|---|---|
| Event title | SLE 2018: 51st Annual Meeting of the Societas Linguistica Europaea |
| Event type | Conference |
| Location | Tallinn, EstoniaShow on map |
| Degree of Recognition | International |