Developing a new perfect: the rise of the Icelandic vera búinn að-perfect

Abstract

Icelandic is basically a “straight-have-perfect” language, where constructions with be + participle are virtually restricted to stative expressions like er farinn “is gone”. But since around 1600, Icelandic has been developing a new perfect consisting of the verb “be” together with the adjective (or participle) búinn + infinitive. The word búinn normally means “finished, done” so it is not surprising that the earliest examples of this construction typically involved transitive telic predicates and animate agentive subjects and had a clear resultative reading. Gradually, the construction developed into a more general perfect, which today can be used with predicates of different types and have a universal and even to some extent experiential (existential) reading. This development is traced and the restrictions on this new perfect in the modern language are described. It is shown that this new perfect has been gaining ground for centuries and is apparently still on the rise, as can be seen from the fact that it is more popular among young speakers than with the older generations. It is also acquired early and apparently more frequent in child language than the standard have-perfect.

Keywords:

• Perfect
• have-perfect
• búinn-perfect
• resultative reading
• universal reading
• experiential/existential reading
• Icelandic

Acknowledgements

Many thanks to the organizers and the audience at the auxiliary symposium at Bernstorff Slot, Denmark, in June 2016 for useful discussion, and to two anonymous reviewers of this paper for valuable comments.

Notes

¹ Note that if oft “often” is left out in (1a), the most natural reading will be inferential: Flugvélin hefur komið of seint = “The plane has apparently arrived late”. Similarly, leaving aldrei “never” out in (1c) would most naturally lead to an inferential interpretation: Jón hefur þvegið sér = “John has apparently washed.” Thanks to a reviewer for pointing this out. See also the discussion around (26)–(27).

² In this paper I will use word-by-word glosses of linguistic examples when they are necessary for the understanding of the form of the relevant constructions. They are typically omitted when the structure of the Icelandic example is very similar to that of its English counterpart or when many similar examples have already been presented.

³ As will be explained below, some linguists use the term experiential about readings that are called existential by others. I will mostly use the term experiential. Kress (1982, 152) uses the German term resultativ-situative Konjugation about constructions with vera + ptcp, indicating that the typical reading is resultative. The most common readings of perfects will be reviewed and illustrated in Section 2.2.

⁴ Interestingly, examples like (3a, b) improve when a temporal specification is added as in (i)‒(ii):

⁵ Whether or not the construction vera + ptcp in Icelandic should be called a perfect is arguably a terminological issue. Sigurðsson (2010, 39, 40), for instance, calls this construction a stative perfect but points out that it can only be used with a limited set of verbs, as already mentioned, and can only have the stative reading. Other researchers seem to consider stative resultatives with “be” + ptcp more like a starting point in the development of true perfects (see e.g., McFadden and Alexiadou 2010 and references cited there). We will briefly return to this question in the conclusion. –Yamaguchi and Pétursson (2003) assume, on the other hand, that vera + ptcp is a true perfect construction in Icelandic and they maintain that it can have an experiential (or existential) reading and not only a stative (or resultative) one as standardly assumed (see also Yamaguchi 2000). The arguments they present for this conclusion are unconvincing, however, since most of the examples that are crucial for their argumentation sound unnatural.

⁶ Following Wide (2002), I will use the item búinn when glossing the new perfect, regardless of the actual form of the word in the relevant example (e.g., fem. búin, neuter búið, etc.). The reason is that glossing búinn as “finished” in such examples is actually misleading. In idiomatic translations into English I will just use a have-perfect and comment on differences in reading when necessary.

⁷ Jónsson (1992, 136) assumes that examples like (6) can only have the universal reading. But consider the following:

⁸ Jakub (1970, 179) claims that the búinn-construction is a marginal lexical construction and not a true perfect. The main reasons for this conclusion are the various restrictions on the búinn-perfect that will be discussed in Section 2.2. Wide also maintains that the búinn-construction “has not yet been fully grammaticalized as a perfect” (2002, 7) although it is “not restricted to a limited set of verbs” (2002, 179). Similarly, Larsson (2008, 87, 88) concludes that it “has not yet developed all the properties of a perfect tense” althouth it “might be about to become a perfect”. The facts to be discussed below should shed some light on this (again, partly terminological) issue.

⁹ Counterfactual constructions in Icelandic are characterized by past tense subjunctive morphology:

¹⁰ All by itself an example like Ég hef fengið flensu is most naturally interpreted as experiential: ‘I have had a flu”. With a context like “I cannot come to the party tonight”, the resultative reading becomes easier to get, especially if one adds the adverb “probably”, as in (7c). Comments from a referee and discussion with Sigríður Magnúsdóttir helped clarify this point.

¹¹ McCawley used the term existential in his influential paper (1971) and it is also the term used by Jónsson (1992), Thráinsson (2007) and Sigurðsson (2010). Here I will mainly use the term experiential about the relevant type of perfect since it seems to be more common in recent literature on tense and aspect (see e.g., Iatridou, Anagnastopoulou, and Izworski 2003; McFadden 2007; Larsson 2008, 2009; McFadden and Alexiadou 2010; see also Comrie 1976). Another reason for this choice of terminology is the fact that the term existential sometimes has a wider meaning in the literature.

¹² Jónsson (1992, 139n) maintains that examples like (13a) have a resultative reading of sorts. Examples beginning with the búinn-perfect Ég veit ekki hvað X er oft búin(n) … “I don’t know how many times X has …”, similar to (13b), are easy to find on the Internet.

¹³ Note that if sofa means something like “take a scheduled nap”, e.g., as babies do, then Barnið er búið að sofa is fine in the sense “The baby has (already) taken his/her nap”.

¹⁴ timarit.is is an on-line corpus containing materials from newspapers and journals in Iceland, the Faroes and Greenland from the nineteenth century onwards.

¹⁵ This example sounds like something you could find in a recipe, as can be verified by a Google search.

¹⁶ Inferential readings of “perfects” seem to occur in a number of languages, whatever the reason may be (see e.g., Jónsson (1992, 137), who cites Dahl (1985)). See also Guéron (2015, 107), who gives the examples in (i) while (ii) is from Rothstein (2005, 5):

¹⁷ Examples of this sort are fine if vera “be” is in the past tense: Hann var búinn að lesa bókina í gær “He had read the book yesterday.”

¹⁸ It is also possible that týnast “get lost” is not immediately compatible with the búinn-perfect because it is an intranstive inherently telic predicate like detta “fall” and deyja “die”, cf. the discussion around (21)–(22).

¹⁹ Originally past participle of the verb búa “prepare”.

²⁰ A quick search in the online version of the Ordbog over det nordiske prosasprog (https://onp.ku.dk/, under the verb búa) did not reveal any Old Norse examples of the búinn-construction in the modern meaning.

²¹ Example (37) is difficult to interpret because the infinitive brenna could either be the weak transitive brenna (past tense brenndi) or the strong intransitive brenna (past tense brann). The weak transitive verb is used in the immediately preceding context (brenndi upp Reiðingahúsið og allt það þar var inni; “burned up the tool shed and everything inside it”) and since the búinn-construction does not go well with intransitive verbs like brenna and other early examples typically involve transitive (and telic) verbs, I am assuming here that brenna in (37) is the weak transitive verb and the structure of the first clause is elliptic (object of brenna omitted).

²² Unfortunately, many of their citations are so short that one has to look the examples up in the original sources to determine the meaning.

²³ This figure is still much lower than typical figures reported by Wide (2002, 172) for the occurrences of the búinn-perfect in the spoken language corpora she studied. In her study the frequency of the búinn-perfect varied considerably depending on the type of spoken language data considered, being proportionately highest for child language data, as will be discussed in Section 4.

²⁴ In addition to (37) these examples are (with the spelling modernized): korn var búið að vinna “one had (already) harvested the grain”, eftir það að búið var að halda kvöldsöng “when one had performed the evening song”.

²⁵ This project was a part of the Scandinavian Dialect Syntax network, cf. https://websim.arkivert.uit.no/scandiasyn/.

²⁶ In this table, token frequency refers to the total number of occurrences of each type of perfect in the relevant corpus and type frequency is the total number of occurrences divided by the number of different verbs that each type of perfect is used with (see Wide 2002, 176).

²⁷ Andrason bases his paper on a corpus of about 500 sentences uttered by adult students that attended courses he gave for workers at three different companies in the fishing industry in Reykjavík (see Andrason 2008, 124). He does not give any information on the speakers’ background. I follow Andrason in glossing the invariant búna as búna to distinguish it from the normally inflected form glossed here as búinn.

²⁸ It should be noted in this connection that since the auxiliary vera “be” is omitted in the búna-perfect, no tense distinctions can be shown.

²⁹ Thanks to a referee for pointing this out to me.

³⁰ Thanks to Peter Juul Nielsen for pointing out a similar example in Finnish. See also Wide (2002, 63) for a Swedish example.

³¹ Wide (2002, 52n) reports, in fact, that something reminiscent of the Icelandic búinn-perfect is found in the Finland Swedish Munsala (Ostrobothnian) dialect , namely ha “have” färdig “ready” + supine, e.g., jag har färdigskådat teve “I have watched TV”.