Hubert Haider: On expletive, semantically void, and absent subjects
What follows expounds empirical and theoretical reasons for a basic, but frequently neglected, differentiation of the subjects of impersonal constructions. An essential difference is the syntactic distinction between a semantically void subject argument on the one hand and an expletive item for an obligatory structural subject position on the other hand. The primary source of evidence will be passive and middle constructions of intransitive verbs in Germanic and Romance languages. The obligatory preverbal subject position of SVO languages is the grammatical source of subject expletives. In the SOV and VSO clause structure, arguably, there is no obligatory VPexternal subject position and therefore no room for expletive subjects. SOV and VSO languages allow for genuinely subjectless clauses. Expletives have to be distinguished from semantically void subject pronouns. The subjects of intransitive middles, for instance, are semantically void arguments of the verb while the structural subject of an intransitive passive, as in French or in Scandinavian languages, is a nonargumental item in a structurally obligatory subject position. Semantically void subjects may be lexical or null, depending on the null-subject property of the given language. Expletive subjects cannot be null. In the linguistic reality, there is no such thing as an “empty expletive”, contrary to widely shared assumptions in the literature. Whenever analyses or descriptions of so-called impersonal constructions confuse the qualities mentioned in the title, they become inconclusive. In the linguistic reality, clauses may be genuinely subjectless or they may contain an expletive subject or a semantically void argumental subject, which, in a null subject language, is phonetically null but syntactically recoverable. What they never contain is a “null expletive subject”. An “empty expletive” is – as shall be demonstrated – a grammatical concept without factual basis. Its motivation rests on an SVObiased perspective on the clause structure of SOV and VSO.
Janayna Carvalho: Incorporated subjects in Existential Impersonal Sentences in Brazilian Portuguese
In this work, I discuss some characteristics of existential null subjects in impersonal sentences in Brazilian Portuguese. I call these constructions existential impersonal sentences (EIS) and argue that their subject is incorporated into the predicate due to their lack of phi-features. To sustain this claim, I discuss several properties of EIS, namely, incompatibility with individuallevel verbs, non-licensing of anaphora, null possessives, subject-oriented adverbs, and scope inertness as well as their syntactic position after movement of the verb.
Thórhallur Eythórsson, Anton Karl Ingason & Einar Freyr Sigurðsson: Flavors of reflexive arguments in Icelandic impersonals
We argue that reflexive pronouns in Icelandic come in two flavors: with and without a D-layer, which we refer to as Weak Explicit Arguments (WEAs) and Strong Explicit Arguments (SEAs). The simplex reflexive pronoun sig of inherently and naturally reflexive verbs is a WEA whereas the complex reflexive pronoun sjálfan sig of naturally disjoint verbs is a SEA. SEAs are, nonetheless, in some cases realized simply as sig, without the intensifier sjálfan. These two types correspond directly to the distinction made in Landau (2010) between Weak Implicit Arguments, which are D-less φPs, and Strong Implicit Arguments, which have a D-layer. We give arguments in favor of our analysis of WEAs vs. SEAs; one crucial difference is that only SEAs can license secondary predicates — under Landau’s (2010) account, this is expected if SEAs, but not WEAs, have a D-layer. Furthermore, we discuss the Reflexive Passive (ReflPass) in Icelandic and argue that it only licenses D-less φ-bundles, that is, WEAs. On our approach, the ReflPass does not have a syntactically projected antecedent of the reflexive pronoun. That creates apparent problems for case and binding. However, we propose that WEA sig is not only a realization of accusative case but also of nominative case. We also argue that the binding facts are not a problem as WEA sig in the passive only requires semantic binding; adopting Legate’s (2014) analysis of the New Impersonal Passive in Icelandic, we argue that WEA sig only restricts an argument position but cannot saturate it. Once Existential Closure applies, the unsaturated internal and the external arguments are bound by the same operator, ensuring identity between the two.
Sigríður Sigurjónsdóttir & Joan Maling: From passive to active: diachronic change in impersonal constructions
Maling/Sigurjónsdóttir (2002) argued that so-called “impersonal passives” are in principle syntactically ambiguous; they can be analyzed either as canonical passives or as impersonal actives containing a null subject. The syntactic behavior of such constructions can change over time as a result of this ambiguity, as shown by the diachrony of the Irish autonomous form, the Polish –no/to construction, and the New Transitive Impersonal (NTI) which is developing in Icelandic. We argue that all three are accusative-assigning participial constructions which developed from canonical passives, but now display the syntactic properties of an active impersonal construction. On the other hand, the syntactic behavior of the Ukrainian –no/to construction, for example, shows that even a construction in which a transitive verb continues to govern an accusative object may be categorized as syntactically passive. In this paper, we focus on the on-going development of the NTI construction in Icelandic. The NTI is a syntactic innovation that surfaced most noticeably in the speech of children and teenagers in the late 20th century and is now rapidly gaining ground. The proper syntactic analysis of the NTI has been the subject of lively debate among linguists. In this paper, we compare its syntactic properties with the Polish –no/to construction, where the reanalysis from a canonical passive to an active impersonal construction has been completed, and review evidence that the NTI developed as an extension of the impersonal passive of intransitive verbs, a construction which is part of the standard language
Anne C. Wolfsgruber: Impersonal interpretations of Medieval Romance se - tracing initial contexts
Passive se and impersonal se constructions have inspired many synchronic syntactic studies. The diachronic path that is involved in the reanalysis of se in that direction has not been of equal interest. The present paper investigates the developments that lead to the reanalysis of se from being a marker of passive constructions towards that of marking impersonal active sentences. The role of se with impersonal verbs as well as se in fixed indefinite chunks is explored and the relationship of these constructions with the more global reanalysis passive > impersonal active is scrutinized. Se co-occurring with impersonal verbs as well as se in fixed indefinite chunks seem to be older constructions that can be linked to the reflexive taking over parts of the functional spectrum of the R-morpheme in Late Latin and they evidence important details on the initial stages of ambiguity in the reanalysis of passive > impersonal active.
Eduardo Amaral & Wiltrud Mihatsch: Incipient impersonal pronouns in colloquial Brazilian Portuguese based on pessoa, pessoal and povo
The best-studied impersonal pronouns derived from lexical items are so-called MAN-impersonals in several Standard Average European languages, most of which began to pronominalize very early on starting in late Antiquity and the Middle Ages. Remarkably, Portuguese, Spanish, Italian and English lost the (partially) pronominalized expressions after the Middle Ages. Today, French on, German man and Mainland Scandinavian man are the most successful survivors. However, notably in Brazilian Portuguese, there is a subsequent wave of lexical items transforming into impersonal pronouns, not only a gente, originally ‘the people’, starting out as an impersonal pronoun in the 18th century and today firmly established as a first person (plural) pronoun, but, more recently, in the 20th century, incipient impersonal pronouns based on general nouns. This is the case of o pessoal originally ‘the staff’, o povo, originally ‘the people (of a nation)’, and a pessoa ‘the person’, uma pessoa ‘a person’ and as pessoas ‘the persons’ (and a few other very colloquial expressions such as galera ‘guys’). We will analyse the referential as well as the morphosyntactic properties of these impersonal expressions and trace the grammaticalization paths, which in the case of a pessoa/uma pessoa neither seem to correspond to the known path of MAN-impersonals, nor to that of a gente, nor to the path of third person plural impersonals. We will close with some typological reflections, trying to relate the gradual and partial loss of the pro-drop status of Brazilian Portuguese to the new wave of lexically based impersonal pronouns.
