Leiv Inge Aa. Norwegian Verb Particles (Studies in Germanic Linguistics 4), Amsterdam/Philadelphia: John Benjamins Publishing Company. 2020. Pp ix + 184.

1. Introduction

Leiv Inge Aa’s book Norwegian Verb Particles adds to the literature on verb particles in Germanic, which is already rich. The following is a sample of book-length treatments: Fraser (1976); Svenonius (1994); den Dikken (1995); Norén (1996); Toivonen (2003); Müller (2002); Dehé (2002); Lüdeling (2001); Zeller (2001); McIntyre (2001); Dehé et al. (2002); Cappelle (2005); Elenbaas (2007); Thim (2012); Los et al. (2012); Walková (2013); Larsen (2014). Den Dikken noted already in his 2002 review of Zeller (2001) that “[t]he morphosyntactic literature is replete with discussions of verb-particle constructions” (p. 145). He also remarks that “[t]he literature on particles, besides lacking consensus on the proper delineation of the set of particles, also reveals a striking lack of agreement on the representation of the morphosyntax of verb-particle constructions — a novice in the world of particles would easily throw up his hands in despair after an initial perusal of the seminal literature, complaining that there are as many analyses of verb-particle constructions as there are publications discussing their morphosyntax” (145-6). Even though it has been 20 years since Den Dikken made these remarks, the situation is much the same today. There is no consensus on how to analyze verbal particles.

Aa’s careful analysis of Norwegian particles helps shed light on the dazzling array of analyses that Den Dikken called attention to. Aa shows that even within Norwegian, verbal particles do not form a uniform class. He specifically argues for three points. First, he shows that there is dialectal variation with respect to the word order. In Norwegian, particles can either precede or follow the direct object, but there is a strong preference for the particle-object order. Second, Aa argues that the ordering of particles is in part dependent on meaning: idiomatic particles are base generated before the object and locative particles after the object. Third, Aa assumes that particles that precede the object are syntactically distinct from the particles that follow the object. He analyzes the former as non-projecting words and the latter as words that project a phrase.

Section 2 below provides a brief summary of the book. Section 3 provides support from across Germanic for two of Aa’s main points: 1) There is more variation with regard to word order than what is typically reported in the literature. 2) The distribution of particles is partly determined by their meaning. Section 4 offers some concluding remarks.

2. Summary

The book is well written and it is also nicely organized. The introductory chapter introduces the main empirical puzzles that Aa considers. Norwegian particles are said to appear optionally on the left or the right of the object, like in English: pick up the garbage or pick the garbage up (see (4) below for a Norwegian example). Aa refers to a particle that appears on the left as an LPrt and a particle on the right as an RPrt. He argues for the following: “The LPrt and RPrt constructions are semantically distinct and the LPrt construction is the unmarked, preferred, and more frequent alternative in Norwegian” (p. 3). After briefly introducing some further types of examples that will be analyzed in the book, the basic theoretical assumptions are established. Aa adopts a basic Principles & Parameters approach in combination with a neo-constructionist exoskeletal framework. Section 1.4 is devoted to a detailed explanation of the methods of data collection. In addition to examples from previous literature and dictionaries, Aa makes use of corpora, acceptability judgements and also introspective examples.

Chapter 2 presents the Norwegian verb-particle data, along with some references to other Scandinavian languages and English. Many examples are drawn from the Nordic Dialect Corpus (Johannessen et al., 2009), and it is nice to see this rich resource being put to use. Aa emphasizes that the oft reported optionality in particle placement is an oversimplification of the Norwegian data. Based on earlier sources (especially Aasen 1848; Sandøy 1976) and fresh corpus and fieldwork results, Aa shows in section 2.1 that the LPrt pattern is strongly preferred, and in some dialects obligatory. Aa remarks that this fact has long been recognized in the descriptive tradition, but is nevertheless often ignored in the theoretical literature. The chapter ends with a list of empirical generalizations drawn from the Norwegian data. As already mentioned above, LPrt and RPrt differ in meaning. Aa also shows that verb-particle combinations form a single prosodic unit for many speakers. Furthermore, ground promoting particles do not allow RPrt, but RPrt is preferred in complex verb particle constructions. In sum, the chapter provides interesting material that relate to a number of different topics. Chapter 2 can be read as a stand-alone chapter by researchers wishing to access the main empirical facts of Norwegian particles without delving into the more theoretically oriented parts of the book.

The third chapter provides an overview of previous approaches to the syntax of verb particle structures. Aa highlights strengths and weaknesses of previous analyses, and he convincingly shows that there is still work to be done. The chapter is organized around two basic questions: First, how should the two possible word orders be analyzed? Is there a ‘basic’ word order, and if so, is it LPrt or RPrt? Second, what is the grammatical status of a particle? Does it head a small clause or does it form a complex predicate with the verb? Aa argues against a uniform analysis: deriving LPrts from RPrts or vice versa does not capture the empirical facts properly. Specifically, such proposals do not capture the Norwegian preference of LPrt over RPrt, nor do they account for the fact that the two differ in meaning. The meaning difference is not completely straightforward to pinpoint, but it is in many cases quite clear. This can be seen in the contrast between examples such as (1) and (2), provided by Aa (page 53):

Aa shows that the main generalization about meaning and word order is that non-spatial particles are LPrts. RPrts tend to be spatial, but spatial particles can also be LPrts.

The review in chapter 3 of previous work unfortunately omits the monographs by Norén (1996), Lüdeling (2001), and Dehé (2002). Dehé is only briefly cited, and Norén and Lüdeling are not even mentioned in passing. These omissions are likely due to the fact that these authors do not focus specifically on Norwegian. The abundance of work on this topic mentioned above is also relevant: at this point, it is probably close to impossible to compose a fully comprehensive review of the published research on verbal particles in Germanic.

Chapter 4 presents Aa’s analysis. Aa adopts Larsen’s (2014) analysis that particles are optionally projecting words and that LPrts fundamentally differ from RPrts in terms of structure. This is also what I have myself proposed in independent work on verbal particles in Swedish (see, e.g., Toivonen 2003). Aa posits the structures in (5)–(7) for the examples in (3)–(4). For optionality, I adopt the same convention as Aa: curly brackets indicate alternation and mutual exclusion. For example, in (4) the particle ut can appear before or after hunden, but not both.

Non-spatial particles appear as non-projecting words adjoined to V⁰, as in (5). The basic structure for spatial particles is shown in (6), but spatial particles can also merge as in (6) and remerge in V⁰, as illustrated in (7). For spatial LPrts, Aa also considers the possibility that they have the structure in (5), like non-spatial particles. He rejects this alternative analysis, mainly because his proposal models the hypothesis that structure reflects meaning: non-projecting particles are non-spatial and projecting particles are spatial. However, this analysis leaves unexplained the fact that spatial particles occur more frequently as LPrts than RPrts, as Aa shows on page 48.

Chapter 4 reaches beyond the distinction between spatial and non-spatial particles. It also covers complex particle constructions (8), unaccusative particle constructions (9), and ground promoting particles (10). Ground promotion in particular has previously not received much attention in the literature on verbal particles, and Aa’s careful consideration and new examples are therefore a welcome contribution. Examples (8)–(10) are from Aa (pages 135, 157, and 147).

Chapter 4 further contributes a discussion of the interaction between structurally determined semantics, the meaning of individual lexical items and interpretation imposed by non-linguistic world knowledge. Aa shows how some of the variation and many subtle judgements can be explained by interactions between the different types of meaning.

Finally, chapter 5 offers a summary, highlighting the main theoretical points argued for in the book.

3. The distribution of verbal particles

Many previous analyses of Norwegian particles are based on the assumption that the distribution of Norwegian particles is simply optional, which Aa shows is an oversimplification. To recap from above: (1) LPrts are generally strongly preferred, and (2) there is a meaning difference between LPrt and RPrt. These important conclusions raise the question of whether we need to revisit the standard generalizations about verbal particles in other Germanic languages as well. I briefly consider each of the two points in this section. Section 3.1 consider the basic word order generalizations across Germanic, and Section 3.2 considers meaning and particle position. I present pointers to previous literature suggesting a richness of the verb particle constructions which can be compared to what Aa finds in Norwegian.

3.1 Dialectal variation in word order

Aa remarks in the first chapter that the Scandinavian languages famously display an interesting typology of particle ordering (Taraldsen, 1983; Herslund, 1984; Svenonius, 1996; Faarlund, 2019, a.o.). The word order differences in particle placement across Scandinavian can, according to claims of “traditional linguistics” (p. 10), be summarized as follows: particles can appear on either side of the object in Norwegian (4), Icelandic (11) and Faroese (12), particles follow the object in Danish (13), and particles precede the object in Swedish (14). ¹

English patterns with Norwegian, Icelandic and Faroese: the particle may precede or follow the object. In German, Dutch and Afrikaans, the particle follows the direct object, as illustrated in the German examples in (16), from Lüdeling (2001):

Yiddish differs from German, Dutch and Afrikaans in that the particle precedes the direct object in V2 structures. In Yiddish non-V2 structures, the particle immediately precedes the verb, which in turn can precede or follow the object (Diesing, 1997; Vikner, 2016).

The basic generalizations about cross-Germanic word order variation in V2 structures are summarized in Table 1. Even though they share a lot in common in form, function, meaning, and use, verbal particles differ across the Germanic languages with respect to position in the clause.

Table 1. Word order variation in Germanic in V2 structure

Aa shows that the Norwegian facts are more nuanced than Table 1 would indicate, and Swedish also shows divergences from the dominant pattern. In addition to the standard LPrt distribution, there are dialects that allow RPrts:

The RPrt option is attested among speakers of some dialects spoken in Finland (especially the south of Finland), and the variant is also found in southern Sweden (Kvist Darnell and Wide, 2002).

There is also variation in Faroese. Aa’s initial description echoes much of the literature and indicates optional word order (see, e.g., Hulthén 1948, Thráinsson 2007, and Dehé 2015). However, referring to Sandøy (1976), Aa points out that this generalization is not quite correct: the RPrt option is strongly preferred. The preference for RPrt in Faroese is confirmed by Lundquist (2020).

Danish has been reported to consistently require the RPrt order. As Herslund (1984) puts it, “only one sequence is permitted in Danish: no matter how long and cumbersome the direct object is, it precedes the particle” (p.40). However, Aa suggests that Danish might allow LPrt, at least colloquially.

In the non-Scandinavian Germanic languages, there is variation with respect to the morphological connection between particles and verbs. Particles can typically not be separated from the verb in OV structures. However, in Dutch (Bennis, 1992; Neeleman and Weerman, 1993; Evers, 2003; Blom, 2005), some varieties of German (Müller, 2002, section 6.1.3.3), and marginally in Afrikaans (Le Roux, 1988), the particle can appear in several possible positions if the clause contains more than one verbal element. Standard German and Afrikaans normally require the particle to appear immediately before the main verb in verb-final structures.

Aa hints (e.g., p. 41, p. 43) that the standardly assumed generalizations about particle positioning should perhaps be revisited also in languages other than Norwegian. The examples cited in this section provides support for this suggestion.

3.2 Interpretation and word order

Aa argues that the distribution of Norwegian particles is partly determined by the interpretation. Some further examples of meaning-dependent ordering of particles in Swedish, English, Faroese and Dutch are provided below.

In Swedish and English, there is a difference in meaning between the LPrt and RPrt version of examples such as (18):

The more natural RPrt version goes against the regular particle placement rule in Swedish. The LPrt version only has the odd interpretation where someone brings the elevator down or dismantles it. I argued in Toivonen (2003) and elsewhere that the difference is due to the fact that the LPrt is necessarily predicated of the direct object. Whether or not that proposal is correct, this class of examples supports Aa’s claim that particle placement can be dependent on interpretation.

Another class of Swedish phrases that violate the standard LPrt word order consists of certain reflexive verbs which require the particle to follow the reflexive object. An example is provided in (19), and other examples include ställa sig upp ‘stand up’ and bryta sig in ‘break in’.

In most dialects of Swedish, the particles must follow the reflexive in these examples. ² Interestingly, however, some Finno-Swedish dialects (mentioned in section 3) that allow the RPrt order permit LPrt in these reflexive examples: %sätta ner sig (Kvist Darnell and Wide, 2002).

The RPrt order is strongly preferred in Faroese (Section 3.1). Sandøy concludes that RPrt is the only possibility for non-idiomatic directional particles in Faroese, and optionality is only possible in fixed expressions where the particle is bleached of semantic content (Sandøy, 1976, Section 3.4). In other words, meaning plays an important role in the particle distribution in Faroese.

Bennis (1991) notes that the distribution of particles in Dutch is dictated in part by the interpretation. Specifically, particles can only appear between the modal and the verb when it denotes a result of the verb, and not when it co-occurs with a stative verb such as wonen ‘to live’ (Bennis, 1991, fn. 2).

4. Concluding remarks

Aa shows in Norwegian Verb Particles that the standard generalization regarding the distribution of verbal particles in Norwegian is an oversimplification: the particle placement is not optional. Aa expresses the hope that his study of Norwegian might inspire reinvestigations of related languages (p. 43). The additional examples provided in section 3 of this review indicate that such reinvestigations might prove fruitful: there is interesting empirical material to explore beyond the standard generalizations adopted in the theoretical literature.

Aa’s book is a very valuable contribution to our field. It will certainly be of interest to anyone who is interested in the syntax and semantics of verbal particles and Norwegian syntax. The different chapters and sections contain enough cross-references, summaries and recap for the reader to be able to focus on certain parts over others. This does not, however, make the book seem repetitive. The book is also a good model for how to collect and present linguistic data from a variety of sources with the goal of constructing a theoretical analysis.