Considerable controversy has recently arisen concerning the validity of the notion of the Sprachbund: Campbell (2006: 2) declares that “linguistic areas boil down merely to a study of local linguistic borrowing and its history, and little else”; Stolz (2006: 36; 45) concludes that Sprachbünde are “projections from the minds of linguists.” He excludes geography and culture as criteria to be used for judging the existence of a Sprachbund, and sees the use of the “isogloss-first” method, using the geolinguistic distribution of a feature to assess the extent of a Sprachbund, as uninformative. Should we, then, as Stolz (2006: 46) recommends, “either strip the term of its unwelcome and much too suggestive connotations or abolish it for good”? Or is there still reason to retain the notion as a descriptor of a real phenomenon? This chapter argues that there is, indeed, value in preserving the concept of the linguistic area, but that it will require substantial updating. What is proposed here is a more dynamic, three-dimensional depiction of the linguistic area, as embodied in the concept of a “Stratified Convergence Zone.”

This chapter begins with a brief synopsis of earlier work on contact and areal linguistics, followed by an overview of new empirical approaches, including the EUROTYP and ensuing projects. Finally, it presents the model of a new view of the Sprachbund, as a Stratified Convergence Zone.

Languages in Contact: Foundations

The study of languages in contact, once relegated to the periphery of linguistic inquiry, has recently come to be recognized as a subdiscipline with robust explanatory power, positioned between the disciplines of historical linguistics and sociolinguistics and firmly linked to empirical methodology. Among the earliest linguistic models to represent contact as significant was Johannes Schmidt's (1872) renowned Wellentheorie ‘Wave Theory,’ alongside the work of Hugo Schuchardt (1883, 1884), which helped establish the study of pidgins and creoles. Kristian Sandfeld's (1930) classic Linguistique balkanique accounted for many of the shared innovations in the Balkans as due to the domination of Byzantium and to the influence of the Greek Orthodox Church. In Italy, “Neolinguisti” like Matteo Bartoli (1925) focused on the geographical distribution of innovations, expanding on the explanatory role of centrality and peripherality. In his analysis of the languages of the Pacific Northwest, Boas (1938) showed that innovations could spread across linguistic and cultural borders without impediment.

The Cambridge Approaches to Language Contact (CALC) series was set up to publish outstanding monographs on language contact, especially by authors who approach their specific subject matter from a diachronic or developmental perspective. Our goal is to integrate the ever-growing scholarship on language diversification (including the development of creoles, pidgins, and indigenized varieties of colonial European languages), bilingual language development, code-switching, and language endangerment. We hope to provide a select forum to scholars who contribute insightfully to understanding language evolution from an interdisciplinary perspective. We favor approaches that highlight the role of ecology and draw inspiration both from the authors’ own fields of specialization and from related research areas in linguistics or other disciplines. Eclecticism is one of our mottoes, as we endeavor to comprehend the complexity of evolutionary processes associated with contact.

We are proud to add to our list Bridget Drinka's Language Contact in Europe: The Periphrastic Perfect through History. Few researchers have undertaken the daunting task that the author has embarked upon in this comprehensive book, namely to track down the origins of the periphrastic perfect construction in Europe (which is quite complex) and to trace its diffusion across the map, while at the same time assessing the role of language contact as an essential element in this development. The book takes a wide-angle perspective, spanning the geographical breadth of Europe – from Portuguese to Finnish, from Icelandic to Bulgarian – and it delves into the 2,500-year-old history of the perfects and resultatives on this continent. Besides large-scale developments, Drinka also focuses on the micro-level responses to these “macro-historical processes”: she documents not only the movement of innovations across a population, such as the spread of HAVE-resultatives from German into the West Slavic languages, but also the role of the individual speaker as the transmitter of change. She presents, for example, the remarkable case of the Aragonese scribe who, in May 1147, switched from Visigothic to Caroline script, signaling not just a change in the scribal tradition but also an entire realignment of cultural allegiance toward trans-Pyrenean norms. This realignment is also reflected in the increased use of BE perfects in this area. The role of sociohistorical events as actuators of change is one of the strengths of this book.

Drinka adduces extensive empirical evidence to support her striking – and undoubtedly controversial – conclusions concerning the origin and spread of the perfects.

In a recent article on language contact, Victor Friedman (2012: 411) observes that, while the lexeme ‘have’ exists across the Slavic languages, possessive constructions used in resultative structures are limited to three geographical areas, all somewhat peripheral, and all adopting the structure through the operation of calquing:

1. • in the Balkan Slavic languages (Bulgarian, Macedonian, and the Torlak dialects of SE Serbia and Southern Kosovo) in contact with the Balkan Romance languages (Romanian, Aromanian, and Meglenoramanian), Albanian, and Greek;

2. • in N. Russian in contact with Finnic; and

3. • in Czech and Polish, in contact with German.

The role of contact in the development of the possessive perfects of the Balkans and of N. Russian has been discussed in some detail; in this chapter, we turn our attention to the development of HAVE perfects in Czech, Polish, and the other W. Slavic languages, in western Ukrainian, and in Lithuanian. What we will discover as we briefly examine each of these varieties in turn is that the W. Slavic languages, especially the vernaculars of Czech and Polish, were influenced directly by Western Europe, and developed resultatives modeled on the perfects of German, while Kashubian and the Baltic Sea Slavic varieties, such as Slovincian and Polabian, were influenced even more thoroughly by German, to the extent that the HAVE and BE constructions may be considered perfects. Lithuanian, on the other hand, presents a unique profile that requires further examination.

HAVE Resultatives in the West Slavic Languages

We turn now to the HAVE + -n/t- construction, identified by Clancy (2010: 160) as the “New Slavic Perfect.”

Czech

As noted previously, Czech has had a HAVE resultative from the beginning of its written tradition.

Centuries-long contact with German fostered its growth in the vernacular language (Dickey 2013: 115 fn), though it came to be suppressed in the standard language (Thomas 2003). As in other W. Slavic languages, the HAVE resultative in Czech was first formed with an objective complement and a PPP agreeing in gender, number, and case, but its later loss of agreement in the participle has led some scholars, such as Dickey (2013), to reclassify Czech's resultative structure as a perfect.

Introduction

In Chapter 1, a stratified depiction of the expansion of the perfect across Western Europe was sketched, and in Chapter 7, the stages of this development were laid out. In this chapter, we focus on the third stage of this chronology:

In Chapter 3, the shift of Anterior > Preterite was examined in light of claims by Bybee, Perkins, and Pagliuca (1994) and others that it represents a universal process, governed by such principles as source determination, unidirectionality, and universal paths (see Figure 3.1 in Chapter 3). In response to those claims, we presented evidence that, rather than being motivated by universal paths, the shift in Europe had a fairly narrow point of initiation, in and around Paris, and that it diffused first to areas within the cultural sphere of that city, and from there to contiguous areas. The similar shifts in French, German, and other varieties should therefore not be regarded as separate instantiations of the shift, but rather as the long-lasting operation of a single, widely diffused shift. This fact is well illustrated by Ternes's (1988: 340) mapping of the contiguity of the languages in the “analytic group”, which have undergone the Anterior > Preterite shift (Figure 10.1). This group includes northern French, southern German, northern Italian, Czech, Slovak, Slovenian, and northern Serbian and Croatian. The area where the Anterior > Preterite shift occurred is clearly visible in Figure 10.2.

This contiguity of the affected languages constitutes strong evidence for areal diffusion as the prime motivation for the spread of the innovation in Europe. The Slavic languages located in the vicinity of the nuclear area – Czech, Slovak, Slovenian, and northern Croatian and Serbian – follow the same pattern in using their analytic perfects (with active resultative l-participle + BE) to mark the preterite.

Introduction

In the present chapter, we turn to Maslov's Stage 2, the development of the l-perfect across the N. Slavic languages. After a discussion of the replacement of the old aspect tenses by l-perfects in N. Slavic, we focus on geographical factors and examine the documentary evidence. An examination of the role of sociolinguistic factors in the development of the l-perfects east and west and the impact of this distribution on the expansion of aspectual contrast concludes this chapter.

Replacement of Aspect Tenses by l-perfect in North Slavic

In Chapter 11, we noted the general tendency toward loss of the imperfect and aorist categories in S. Slavic, especially in those varieties at a greater distance from Greek. This trend is likewise to be noted in the N. Slavic languages: most varieties lose the imperfect first, followed by the aorist, replacing them with the l-perfect, which becomes the only marker of the past. Those languages with more conservative traditions, such as Czech, tend to maintain the aspect tenses longer; those with more innovative tendencies, such as the northern Russian dialects around Novgorod and Pskov, lose the imperfect and aorist very early: there is, for example, practically no evidence for either of these categories in the vernacular birchbark letters of Novgorod. The imperfects and aorists survive in Old Church Slavonic and in medieval Russian and Croatian Church Slavonic, and, as we have seen, in the modern S. Slavic languages of Bulgarian, Macedonian, and Serbian. The aorists and imperfects are already almost extinct in the earliest documents of Old Polish and Old Czech (Andersen 2009: 127). These facts are summarized in Table 13.1.

We can note two conservatizing forces at work in Slavic, which hindered the process of replacement:

1. 1) the archaizing pressures of Church Slavonic, and, ultimately, Byzantine Greek, which affected the S. Slavic languages most thoroughly, but which had an ongoing resurgent effect in all areas under the influence of Orthodoxy (Chapter 12)

2. 2) the retention and revamping of old categories for stylistic purposes (Czech) or in imitation of external models (Sorbian), which Dickey (2011: 201) calls “replica preservation” (Section 13.6)

Chapter 13 focused on Maslov's second layer, the distribution of the l-perfect in N. Slavic, the loss of anterior meaning, and the development of preterital value for this construction across the N. Slavic languages. In the next two chapters, we turn our attention to Maslov's third layer, exploring the introduction of new periphrastic perfects and resultatives in *ṷes- or *–n/t- in the N. Slavic and Baltic languages. The present chapter begins with an overview of the distribution of these structures, and proceeds to an examination of the role of contact in the development of each. With regard specifically to the Circum-Baltic region, we follow Stolz (1991), Nau (1996), and Koptjevskaja-Tamm and Wälchli (2001) in endorsing the modification of the concept of Sprachbund for this area; what is envisioned here instead is the notion of a “Stratified Convergence Zone,” a three-dimensional model depicting layers of contact across space and time. Evidence from the new resultatives of Baltic and N. Slavic, especially in the Circum-Baltic area, strongly supports the validity of this proposal: both the *ṷes- and the *–n/t- resultatives show signs of stratified development and diffusion at the micro- and macro-level.

The “New Resultatives” of North Slavic and Baltic

In their comprehensive study of the resultative structures of the N. Slavic and Baltic languages, Wiemer and Giger (2005) explore the present-day distribution of the periphrastic resultative structures, especially those constructed with participles in –n/t- or -(v)ši /-vęs, with an eye to areal spread and the effects of grammaticalization. We will make full use of their findings in this section as we attempt to sort out the historical foundations for the current, complex distribution of these structures.

As the l-perfects took on preterital value in virtually all the N. Slavic languages, other participial structures came to be used more productively to fill various functions, whether as subject-oriented resultatives, object-oriented resultatives, or possessive resultatives.

Core vs. Periphery in the Charlemagne Sprachbund

In the conclusion to his edited volume in the EUROTYP series on adverbs, Johan van der Auwera (1998b: 824) introduces the term Charlemagne Sprachbund to refer to the area of Western Europe where languages adhere to an array of similar grammatical patterns and thus provide evidence for the existence of a European linguistic area, also called Standard Average European (see Chapter 2 for discussion). Van der Auwera notes that the “core” areas of the Sprachbund are French, German, and Dutch, with Italian and Polish closely linked – all located within or near to the area ruled by Charlemagne. He demonstrates the degree of similarity that exists across European adverbial constructions by assembling a list of the features that tend to occur more frequently in the core area and less frequently in the periphery, counting the occurrence of these features, and grouping the varieties with similar numbers of features on an isopleth map (Figure 7.1) (cf. Haspelmath's similar map, presented as Figure 2.1 in subsection 2.3.1). For reference, we can examine Figure 7.2, which is an actual map of the holdings of Charlemagne at the time of his death in 814.

Van der Auwera (1998: 824) gives the following explanation for this distribution:

He thus points to the essential role of both centrality and peripherality as explanatory principles of innovation through contact, since the profile of groups who did not adopt a change can be as telling as information about those who do.

Van der Auwera goes on to present a similar mapping of temporal-aspectual features of the languages of Europe, based on Thieroff (2000: 265–305), which also points to French and German as core members of the Sprachbund (Figure 7.3).

The features that van der Auwera (1998: 833 n13) utilizes from Theiroff (2000) in order to construct this map are shown in Table 7.1

Introduction

In the previous chapter, our focus on the historical interactions of the Balkan languages, including the South Slavic varieties, revealed the preeminence of Greek in that area and the essential role language contact played in the history of the perfects there. In the present chapter, we focus more particularly on the influential status of Byzantine Greek as emblem of the Byzantine Empire and the Greek Orthodox Church and on the “roofing” effect that this influence had on the perfects of Old Church Slavonic. We construct a chronologization reminiscent of that formulated by Maslov (1988: 70–85) in his classic work on the typology of the resultative, the perfect, and aspect. Maslov distinguishes three layers in the development of resultatives and perfects in the history of Slavic:

1. 1) the Indo-European layer, with its synthetic stative perfect

2. 2) the expansion of the perfect through the introduction of analyticity in such forms as the l-perfect, and

3. 3) the creation of new possessive resultatives and perfects built on past passive participles in -n/t- and of non-possessive structures such as that built on the gerund in –(v)ši.

Having discussed the Indo-European perfect in detail in Chapter 4, we begin here with an examination of the second layer, and continue this discussion into the next chapter; the third layer will be examined in Chapters 14–15. Throughout this chapter and the next three, we will attempt to sort out the possible motivations for these developments and their complex interactions.

Slavia Orthodoxa and Slavia Romana

In his study of Church Slavonic and the impact that confessional allegiance had on the development of the Slavic languages, Riccardo Picchio (1980: 22) coined the terms “Slavia Orthodoxa” and “Slavia Romana” to refer to the cultural and linguistic divide that existed between those Slavic lands that adhered originally to Greek Orthodoxy and those that practiced Roman Catholicism, respectively. Slavia Orthodoxa, comprising modern-day Serbia, Bulgaria, Macedonia, Russia, Ukraine, and Belarus, remained historically under the influence of Byzantium and the Greek Orthodox Church; Slavia Romana, consisting of present-day Poland, Czechia, Slovakia, Slovenia, Croatia, and Lusatia in Germany, was historically Roman Catholic, and, later, also Protestant. One sign of the contrast between east and west is the use of the Cyrillic and, historically, Glagolitic alphabets in the east and the Roman alphabet in the west – iconic symbols of adherence to different cultural norms and scholarly traditions.

In Chapter 5, we traced the complex history of the periphrastic perfect and related structures in Ancient and Hellenistic Greek. After an overview of the synthetic and periphrastic perfect of Latin, we examine recent theories as to the origin and development of the periphrastic perfect. Among the key claims to be explored is whether the periphrastic perfect of Latin developed completely independently from that of Greek, as is usually claimed, or whether this structure owes at least some of its existence to contact with Greek. We then turn to the evidence for the later influence that Latin exerted on the Greek periphrastic perfects during the time of Roman dominance in Greece referred to in the last chapter (subsection 5.3.4). Finally, we explore in some detail the role that the “sacral stamp of Greek” played in influencing the language of Latin authors in the Christian tradition.

Overview of the History of the Latin Perfect

Latin Synthetic Perfects

As discussed briefly in Chapter 4, the western Indo-European language families of Celtic, Germanic, and Italic all merged the ancient perfect and aorist categories into a single preterital category. Remnants of the ancient reduplicated perfect persist in a few Old Irish, Gothic, and Latin forms, and Germanic strong preterites retain the o-grade of the ancient IE perfects. But in each of these families, the IE perfect no longer exists as a separate category. In Latin, the term perfectum refers to this amalgamated category of perfect and aorist, so that the semantic value of the Latin “perfect” is, above all, preterital, with possible reference to an anterior meaning in appropriate contexts.

Latin Participles

As discussed in Chapter 4, the Latin participial system is very limited compared to that of Greek. It only has a present active, a past passive, a future active, and a future passive participle:

Latin Periphrastic Perfects

Besides the archaic and innovative synthetic perfects, Latin also formed a periphrastic perfect for its passive and deponent verbs. Deponents, which originated as ancient middles (Ernout 1909: 4; Flobert 1975; Tuttle 1986: 247–48), are usually characterized as active verbs with passive morphology e.g., sequor ‘follow,’ cōnor ‘try,’ orior ‘rise,’ queror ‘complain.’

Introduction

In her important work on the development of aspectual systems in English, Laurel Brinton (1988) notes the similarities of periphrastic perfects across the Germanic languages and makes the following claim about their origin:

Brinton concludes that auxiliaries in general, and perfect auxiliaries among them, must have arisen in Proto-Germanic times. It will be the focus of this chapter to explore the key issues inherent in this claim: How indeed can we explain the similarities in the periphrastic perfects across many Germanic languages if not by shared inheritance? Are the features that the Germanic languages share only explainable through the notion of genetic relationship? While I am completely sympathetic with Brinton's unwillingness to posit parallel development of identical systems as an explanation, I propose an alternative explanation to her reconstruction of the periphrastic perfects for Proto-Germanic, namely that the Germanic varieties that developed a periphrastic perfect resemble each other so fully because they followed the same model – that of Latin. In the case of Gothic, the oldest well-attested Germanic language, we will find that Greek played the role as “roof” language more extensively than Latin did, but we will also discover that fascinating evidence for the mutual influences of Gothic and Latin exists, as well. What will emerge from our examination here is that the morphosyntactic patterns of Latin served as the template on which the Germanic languages of Western Europe built their perfects.

The dispute as to the origin of Germanic perfects is an old one: in the Grimms’ Deutsches Wörterbuch, Latin is mentioned as the model for the HAVE perfects of Germanic (DWB Bd. 10 Sp. 71). Meillet (1970) likewise suggests that the Germanic languages developed their HAVE perfects based on the influence of Latin: “An important procedure, not yet utilized by Gothic and doubtlessly owing in the beginning to imitation of Vulgar Latin models, is that which consists of uniting the participle with the verb ‘to have’” (Meillet 1970: 70).

In his insightful analysis of the perfects of Europe, Jouko Lindstedt (2000: 368) remarks that perfects in the languages of the world are usually periphrastic in form, but that a notable exception to this tendency exists in the old perfects of Indo-European, especially as represented by Classical Greek and Sanskrit. To this observation can be added the fact that many Indo-European languages have developed a periphrastic perfect, either as a direct replacement of the IE synthetic perfect or as a complementary addition to the verb system. Why should it be that the old, synthetic perfect of IE came to be replaced by analytic structures in so many IE languages? Why did these periphrastic structures tend to be formed with stative auxiliaries and past passive participles, as illustrated in Chapter 3, at a rate much higher than that found for perfects of other languages? What role did Proto-Indo-European, the earliest reconstructable ancestor of virtually all European languages, play in this rise of the periphrastic perfects of Europe?

In previous chapters, it was claimed that the configuration of components available to speakers of Indo-European languages enabled and even encouraged this development. In Chapter 2, it was hypothesized that the development of the meaning of later periphrastic constructions – resultatives, passives, and perfects – was set in motion by the passive and anterior value of participles and verbal adjectives inherited from the proto-language. In Chapter 3, it was concluded that the use of HAVE as an auxiliary was surely encouraged by the presence of the participle or verbal adjective inherited from PIE. In this and in following chapters, we will expand on these claims, exploring the semantic value and formal features of the oldest perfects in an attempt to discover whether the structures of PIE actually did predetermine the outcome of periphrastic structures in Europe and the extent to which this development was influenced by the interplay of auxiliary and participle.

The Balkan languages, especially Greek, Bulgarian, Macedonian, Romanian, and Albanian, present a remarkable series of similarities in their syntactic structure, in spite of the fact that most of these languages are only distantly related. As mentioned in Chapter 2, the Balkan languages have come to be recognized as a classic example of a Sprachbund, or linguistic area, which can be defined as a group of contiguous languages that share a set of structural correspondences not inherited from a common ancestor but developed by means of long-term contact. What this chapter will demonstrate most clearly is that a Sprachbund is not merely an area of single-tiered contact or a region where features converge in some unmotivated way. Rather, it represents an accumulation of the effects of many layers of influence, all governed by the sociolinguistic pressures that were in existence at the time of the spread of each innovation, as well as by the formal characteristics of the model and the copying varieties. The way that the innovation is inserted into the linguistic system depends on the formal context, but the motivation for the change itself springs from sociolinguistic considerations (Johanson 1992; Bisang 1996). The innovations tend to enter as productive reworkings, renewals, of elements already in existence in the language, resulting in what could be regarded as motivated, sociolinguistically governed exaptations – exaptations not in the sense that the reanalyzed material was useless before, but in the sense that it receives a new, heightened value, imbued as it is with desirable external connections.

Historical Factors Fostering Balkan Linguistic Unity

Before we begin our analysis of the development of the periphrastic perfect in the Balkans, a brief look at the social and political factors that unified the area throughout its history may be useful, for it is clearly the complex shared history of the Balkan speech communities that caused them to develop so many systematic correspondences. One of the most important factors in this development was a shared history of domination – by the Roman Empire, by Byzantine Greece and the Greek Orthodox Church, and eventually by the Ottoman Turks. The various invasions of Byzantine territory by Slavic and Avar tribes in the sixth century led to the weakening of central administration and cities, with “radical ruralization of society” (Fine 1983; Weithmann 1994; Lindstedt 2000: 240).

How universal is the PERFECT category? Should we, with Bybee and Dahl (1989) and Bybee, Perkins, and Pagliuca (1994), call this category a “gram,” a basic construct found widely in the languages of the world, expressing a fundamental semantic value? Or have we created an expectation that such a category exists because it is part of the temporal-aspectual repertoire of European languages? Are there predictable paths of grammaticalization that perfects tend to follow, or should these be regarded as tendencies, to be triggered by more immediate motivations such as sociolinguistic pressures?

Jacob (1994: 62), in his analysis of the perfects of Europe and of the languages of the world, suggests that it is the inner consistency (“innere Kohärenz”) of the formation of this construction across time and space, especially its persistently recurring uniformity (“immerwiederkehrende Gleichförmigkeit”) in various epochs – in Hittite as well as in Late Latin and Romance – that demonstrates that universal, cyclical processes are in operation and that language contact is thus not a prime factor in the development of the perfects of Europe. To what extent is Jacob justified in claiming “innere Kohärenz” in the formation of the periphrastic perfects of the languages of the world, across time and space? In this chapter, we attempt to answer this question by examining the perfect category from a cross-linguistic perspective, to determine whether it should be regarded as universal or not. In the course of this examination, we assess a number of claims concerning the formation and grammaticalization of the category, especially with regard to the “hodological” (“path-oriented”) approach of studies like that of Bybee et al. (1994).

What the arguments and evidence presented here will reveal is that the perfect should not, in fact, be regarded as a semantically unified, universal category, but that it is better seen as representing an array of related semantic features (CURRENT RELEVANCE, RESULTATITIVE, COMPLETIVE, etc.), available for incorporation into language-specific morphosyntactic configurations. With Anderson (1982), Li, Thompson, and Thompson (1982), Bisang (2004), Wiemer (2004), McFadden and Alexiadou (2010), and others, I claim that it is not so much the meta-category of PERFECT that should be regarded as universal, but rather that it is the more minute, closely related semantic or pragmatic properties that speakers use to construct perfects and related structures according to the cognitive and social pressures they encounter that should be so viewed.