This volume brings together twelve original articles on adverbials in Hungarian, the prime outcome of a three-year project at the Research Institute for Linguistics of the Hungarian Academy of Sciences (financed by the Hungarian research found, OTKA). The book's main goal is to investigate the syntactic and semantic behaviour of adverbs and adverbial adjunct constituents. In line with the book's title (and the series in which it appears), special attention is dedicated to syntax and its interfaces with PF (Phonetic Form) and LF (Logical Form). The majority of the articles in this volume specify the extent to which the syntactic distribution of adverbs is determined by the requirements imposed upon syntax by demands of semantics and, in a few cases, by prosody.
In my view, the book succeeds excellently in covering the entire descriptive array of facts about adverbs and a sizeable portion of other adverbial expressions, together with numerous aspects of their theoretical analysis. The volume presents the reader with a comprehensive (and up-to-date) view of the topic, which is especially welcome in the light of the fact that very little has been published in this domain to date. The rounded picture that emerges is also due to the uniformity of theoretical assumptions adopted throughout: most of the articles work with the exact same set of background assumptions concerning the syntactic structure of the Hungarian clause and the theory of adverbial modification, which is very fortunate and reader-friendly when it comes to a multi-authored book of this size. The most important of these shared assumptions, namely the theoretical framework of Ernst (Reference Ernst2002) (as opposed to that of Cinque Reference Cinque1999) for adverbs, is spelled out in the editor's ‘Introduction’. In Cinque's theory, adverbs are universally specifiers in a rigidly ordered set of functional heads, while in Ernst's approach, adverbs merge to positions that match their (lexically specified) selectional needs – the latter can refer to events, propositions, times, and predicates. The order of adverbials that adjoin at the same level is determined via a calculus that is based on a hierarchy of events, propositions and speech acts. The calculus allows any type to be freely converted to any higher type, such that an event can be converted to a proposition, or a proposition can be converted to a speech act (but not vice versa). This calculus makes the right predictions for the ordering of Hungarian adverbs. Consider the order of modal and manner adverbs in (1).
- (1)
(a) [PredPValószínűleg [PredPhangosan [PredPhorkolvalaki]]].probably loudly snores somebody‘Probably somebody is loudly snoring.’
(b) *[PredPHangosan[PredPvalószínüleg [PredPhorkolvalaki]]].loudlyprobablysnoressomebody
The most deeply embedded Pred(icate)P(hrase) is an event, which provides a suitable host for the manner adverb (‘loudly’). The selectional requirement of the modal adverb (‘probably’), however, is different: it requires a proposition, which it receives after the event is freely converted into a proposition. The opposite order is ruled out because propositions cannot be converted to events, according to Ernst's calculus.
As É. Kiss points out, Ernst's adjunction model is completely in line with the scope-driven nature of the line-up of preverbal elements in Hungarian (scopal elements c-command their scope), if we think about selectional requirements similarly to scope. Furthermore, Ernst's theory is also compatible with the distribution of adverbs in the postverbal domain. In Hungarian all adverbs can occur either pre- or postverbally, with identical scope possibilities, and this is easily accounted for in an adjunction framework (provided one allows for right-adjunction), but does not straightforwardly follow from Cinque's theory. For these reasons, Ernst's theory is chosen as the overall framework in the whole volume. I think the choice is justified, yet, I would have liked to see a more detailed discussion (maybe in the form of an entire article) of the (in)compatibility of Cinque's theory with the Hungarian data.
Adverbs and adverbial adjuncts at the interfaces contains a wealth of valuable material, but space restrictions prevent me from entering into a thorough critique of every contribution. I will therefore summarize in detail only those articles that deal with the basic distribution of adverbs in Hungarian. This is a somewhat arbitrary choice, but one that might be useful to those who plan to read the book without any background in the literature on Hungarian adverb placement. In the second part of this review, I will sketchily outline the contents of the rest of the book.
The basic description of the distribution of adverbs in Hungarian is covered in Katalin É. Kiss's opening article, ‘Syntactic, semantic, and prosodic factors determining the position of adverbial adjuncts’. Special attention is given here to the different distribution of Hungarian adverbs in the pre- and postverbal domains. As mentioned above, adverbs show distinct behaviour in these two domains. While they observe fixed ordering restrictions in the preverbal domain (which follows from Ernst's account, and is similar to the orders proposed by Cinque Reference Cinque1999), they exhibit free ordering in the postverbal domain. In the preverbal domain adverbials carry main stress, and frequency, manner and degree adverbs, for example, must be placed in the order of frequency>manner>degree. In the postverbal domain, on the other hand, any ordering of these adverbs is possible. The explanation of the difference follows from the model of É. Kiss (Reference É. Kiss2008), which was designed to handle Hungarian postverbal word order more generally. Besides allowing for right adjunction, the key ingredient of this analysis is the assumption that the overt position of the verb determines a phase, the domain of which (which includes all postverbal constituents) undergoes what É. Kiss refers to as ‘flattening’, a powerful process with the effect of free linearization at PF.
While É. Kiss's study addresses adverb placement in general, ‘Adverbial (dis)ambiguities: Syntactic and prosodic features of ambiguous predicational adverbs’, by Barbara Egedi, turns to a smaller set of adverbs, namely those whose meaning and prosodic expression differ depending on their adjunction site; for example, adverbs that can have either a manner reading or, when attached higher in the phrase structure, a clausal (e.g. subject-oriented) reading. As manner adverbs in the preverbal domain, they adjoin to the predicate phrase of neutral sentences (PredP) and receive stress. As subject-oriented ones, they adjoin outside the main predication of the clause, to the SpeakerDeixis phrase, and remain unstressed. The situation is different in the postverbal domain: ambiguous adverbs can occur in any order, and, when found in a prosodically neutralized environment (e.g. following a preverbal focus), they do not carry stress. The default interpretation in these contexts is that of manner adverb. The subject-oriented reading is available as a marked option, only if it is coupled with a prosodic pattern in which the adverb is flagged by pauses on both sides, clearly placing the adverb into an intonational phrase (IP) on its own. According to the author, subject-oriented adverbs always constitute an IP on their own, but in contexts where their meaning is lexically fixed (i.e. where they are not ambiguous), they can be integrated into the IP of a larger constituent via a PF-rule called ‘IP restructuring’. The fact that ambiguous adverbs in the postverbal domain cannot be integrated in this way and need to retain the IP on their own can be construed as evidence for the fact that PF and LF interface directly (and not via syntax, since the prosodic effects here do not correspond to varying adjunction sites, as the author argues). This is an interesting conclusion that the author unfortunately leaves without much further comment, no doubt due to the fact that the workings of the PF-interface are quite ill-understood. It is not clear, for example, how PF operations are ordered, especially with respect to free linearization (which is the key ingredient for deriving the postverbal orders in É. Kiss's theory). Egedi states that IP restructuring takes place before the free linearization of postverbal elements, but one might ask if it would not make more sense to assume restructuring after the order of constituents is fixed once and for all.
Hungarian adverbs can also be three-way ambiguous, as Egedi shows with the adverb biztosan ‘certainly’, which can be either a manner adverb, or a clausal adverb expressing probability or certainty. On the ‘certainty’ reading, biztosan, like other epistemic adverbs in Hungarian, has somewhat unusual properties. It obligatorily bears focus stress and can occur in questions (which other speaker-oriented adverbs cannot do). Egedi accounts for this by claiming that this type of adverb modifies a so-called ‘verum’ focus projection in the clause.
In ‘Aspect and adverb interpretation – the case of quickly’, Boldizsár Eszes devotes an entire article to another three-way ambiguous adverb, viz. gyorsan ‘quickly’, which can have a manner, rate and aspectual interpretation. It is shown that in line with the specific semantic requirement of each reading, the actual interpretation of gyorsan is jointly determined by the aspectual type of the predicate that is modified by the adverb and the position in which the adverb is merged.
In addition to adverbs of certainty, there is another set of adverbs associated with focus, whose positioning is usually seen as an interesting quirk of Hungarian syntax: negative adverbs of degree, manner and frequency obligatorily occur in the preverbal focus position, thereby occupying a different syntactic position from their positive counterparts. This quirk is explained away in Katalin É. Kiss's second contribution to the volume, ‘Scalar adverbs in and out of focus’. The idea is that focal placement is due to a semantic property of these adverbs, more specifically, the fact that they have a scalar meaning. Scalar expressions do not have a fixed meaning when it comes to the value expressed on the scale, with the result that a scalar expression n can be freely interpreted as ‘n or more’ or ‘at least n’. Hungarian negative adverbs are furthermore defined on a bidirectional scale that has both a negative and a positive domain. For scalar elements whose meaning falls in the negative domain of the scale, the ‘n or more’ interpretation should be blocked, otherwise their interpretation would give rise to an anomaly in the semantics. For this reason, they have to be focused, and thereby assigned an exhaustive reading (‘precisely n’). É. Kiss does not devote much attention in this article to postverbal occurrences of these adverbs, which always have to be post-focal as well (in the sense that they have to be preceded by a preverbal focus). The author claims that postverbal negative adverbs are interpreted as focused, which is an intuition that I do not share.
Another interesting set of adverbs that can be considered the opposite of negative ones is mentioned in this chapter: positive maximal degree and intensity adverbs, such as nagyon ‘very.much’, alaposan ‘thoroughly’, and teljesen ‘completely’. These adverbs cannot be focused under any circumstances; nevertheless in the preverbal domain they immediately precede the (uninverted) verb. É. Kiss argues that these adverbs are adjoined to PredP, but leaves their incompatibility with the focus position unexplained:
- (2)
(a) [TopPJános [PredPnagyon [PredPelfáradt]]].Jánosvery.muchpreverb.tired.3sg‘János got tired very much.’
(b) *JánosNAGYONfáradtel.Jánosvery.muchtired.3sgpreverb
Beyond investigations of adverb placement, the volume contains studies of several other issues related to adverbs and adverbial constructions. In the remainder of this review, the reader finds a brief summary of these contributions.
Edit Kádár's ‘Adverbial versus adjectival constructions with BE’ (listed in the table of contents as ‘Adverbial versus adjectival constructions with BE and the category Adv’) considers the category of adverbs in Hungarian, concluding on the basis of morphological and historic data that they are prepositional phrases (PPs). She also identifies the syntactic position of adverbs used as predicates and accompanied by the verb ‘be’ (as in He is well).
In ‘Incorporated locative adverbials in Hungarian’ (listed in the table of contents as ‘Locative particle and adverbial incorporation at the interfaces’), Balázs Surányi investigates syntactic and semantic conditions under which pronominal locative PPs can occur incorporated into the verb as a preverbal modifier, while at the same time being doubled by a lexical PP, as in (3).
(3) János[PPhozzá]vágtaatányért[PPMarihoz].Jánosto.3sgthrew.3sgtheboard.accMari.to‘János threw the board at Mary.’
The two PPs are argued to be links of one and the same syntactic chain, which the author refers to as ‘light-headed’, since it spells out the head link with a default pro nominal.
Márta Peredy's article, ‘Obligatory adjuncts licensing Definiteness Effect constructions’, focuses on the argument/adjunct distinction, examining obligatory adjuncts that occur in existential constructions with a non-specific indefinite theme (e.g. Ütöttem egy tojást *(a serpenyőbe) ‘I cracked an egg *(into the pan)’). She explains the obligatoriness of these adjuncts adopting Parson's (Reference Parson1990) framework of event semantics and concludes that obligatoriness is not a matter of argumenthood but rather is due to these adjuncts contributing to the non-specific interpretation of the theme.
In ‘Comitative adjunts: Appositives and non-appositives’, Éva Dékány analyses comitative adjuncts, which can modify plural pronouns in Hungarian. These adjuncts result in inclusive meanings of the sort illustrated in (4).
(4) Mianyávalitthonmaradunk.wemother.withhomestay.3pl‘Mother and me stay at home.’
Dékány's study treats these adjuncts as appositive modifiers of the pronoun with which they occur. The pronoun is shown to be internally complex, containing a variable bound by the comitative phrase.
Two articles in the book examine the syntactic derivation of adverbial clauses. Huba Bartos's ‘The syntax of -vA adverbial participles: A single affix with variable merge-in locations’ offers a uniform analysis for all types of adverbial clauses formed by affixing -vA to verbs. The author convincingly argues that all adverbial clause types can be derived in an antilexicalist framework if one assumes variable merge-in points for the -vA morpheme. The attachment site determines the size of the participial clause, which in turn determines additional properties like the meaning of the adverbial clause, the number of arguments in it, and the height of attachment of the adverbial in the matrix clause.
The other article on adverbial clauses, ‘Temporal adverbial clauses with or without operator movement’, by Barbara Ürögdi, gives an account of the internal syntax of temporal adverbial clauses. Ürögdi focuses on the difference between regular temporal relative clauses (like ‘when’-clauses) and those in which an event is relativized. She shows that temporal markers óta ‘since’ and -ig ‘until’ exhibit dual behaviour and are used with either strategy of relative clause formation due to the fact that they can be construed with a punctual or with a durative interpretation.
Ferenc Kiefer's article, ‘Types of temporal adverbials and the fine structure of events’, identifies the maximal number of verb classes that can be distinguished on the basis of combination with temporal adverbials. Using five adverbials (két órán át ‘for two hours, as a time span adverbial’, két óráig óráig ‘in two hours’, két órakor ‘at two o'clock’, két órára ‘two hours long, denoting length of a resulting state’, and két óráig óráig ‘until two o'clock’), he defines nine distinct verb classes, and thereby develops a more fine-grained event structure for these predicates.
Given the nature of Hungarian as a focus-prominent language, it is not surprising to find that the volume contains two articles concerning the interaction between focusing and adverb interpretation: É. Kiss's contribution on obligatorily focused adverbs, and Ágnes Bende-Farkas's ‘Adverbs of quantification, it-clefts and Hungarian focus’. Bende-Farkas's contribution shows that semantic partition (the division into restrictor and scope) depends on the placement of the adverb with respect to focus. The focus (or cleft) constituent corresponds to the scope of a quantificational adverb only when the adverb c-commands the focus, but not when the adverb is c-commanded by the focus. This has consequences for both the semantic treatment of adverbs and focus.
To conclude, this book is a useful collection of recent work, which impresses with thorough empirical coverage and many innovative ideas. The findings and theoretical implications of these studies are relevant to anyone interested in Hungarian and/or Universal Grammar.
