This monograph is the first large-scale corpus-based diachronic study of external possession (EP) constructions in the English language. The study focuses on structures like (1), in which an NP in the dative case (dem Kind / me, the possessor), which behaves as an external argument of the clause, establishes some semantic relationship of possession with another NP in that clause (die Haare / el dedo, the possessum). This kind of structure was available in the Old and Middle English periods, but is preserved in Present-day English only in a few set phrases and expressions of the type look someone in the eye, having overall been replaced by internal possession (IP) structures containing a premodifier in the genitive case, like those in the idiomatic translations in (1a, b).

A characteristic of dative external possessors (DEP henceforth) is that they systematically display a strict affectedness condition, i.e. they are necessarily associated with beneficial or negative effects on the possessor (Haspelmath Reference Haspelmath, Payne and Barshi1999: 111). From a typological-areal perspective, DEPs are considered as prototypically European, and external possession structures of this type are found in languages such as French, Spanish, German, Greek or Russian, to name a few.

Allen's study seriously challenges two widely accepted hypotheses concerning the history of DEP structures in English and their demise over time: the strict connection between the disappearance of DEPs in English and the loss of the dative/accusative distinction (Ahlgren Reference Ahlgren1946), and the impact of the contact with the Celtic linguistic substratum in accelerating (or actually triggering) the change after the Germanic invasions. Allen does so in a very convincing way, combining – as is usual in her – formal linguistic analysis, thorough corpus research and absolute philological rigour. In this respect, one of the many merits of the text is the author's design of her research plan, which takes into consideration variables which had been consistently overlooked by previous researchers in the analysis of DEPs in English, and which eventually turned out to be decisive ones in the elucidation of the problem. Some of these are (i) making a systematic distinction in the grammatical nature of the possessum – whether it was a direct argument of the verb (an NP subject or object) or the object of a preposition (PObj henceforth); (ii) making a systematic comparison of the relative incidence of DEP structures and Internal Possession (IP) ones within the same text or text type; and (iii) clearly distinguishing poetry from prose texts in the analysis. Another problematic aspect of former studies of DEP in the history of English is that Old English was generally regarded as a homogeneous period in the treatment of this phenomenon, the decline and later obsolescence of DEPs being considered a strictly Middle English issue. Allen demonstrates, however, that the decline of external possession in English was gradual, starting off probably even before Old English times and being well under way in that period. This wide-ranging conclusion can only be arrived at thanks to the fine-grid analysis the author does of the impact of DEP and IP structures in an extensive corpus of Old English containing both poetic and prose texts which ranges temporally (in the prose texts) from the ninth to the eleventh century. To this corpus, another section for Middle English is added, which extends from before the m1 period of the Helsinki Corpus (1150–1250) to the end of the m2 period (1250–1350). Finally, Allen provides the reader with a well-balanced analysis of the different sociolinguistic theories based on language contact in the loss of external possession structures in English and the reinforcement of the IP construction. These include theories based on contact with Old Norse populators, the problem of the possible impact of Latin originals on Old English translated texts and – most importantly – the widely accepted hypothesis (Vennemann Reference Vennemann, Lucas and Lucas2002; Filppula Reference Filppula, Filppula, Klemola and Paulasto2008; Hickey Reference Hickey, Nevalainen and Traugott2012) that the loss of the DEP construction in English and the reinforcement of the IP alternative was due to contact on the part of the Anglo-Saxon invaders with indigenous Celts, whose language lacked external possessors. According to the proponents of this theory, the seemingly puzzling fact that no written evidence of such external influence should be available until the Middle English period (the moment when the change from DEP to IP revealed itself in the written records) would be explained by assuming the tenet of the so-called Celtic Hypothesis, which postulates that it was only after the Norman Conquest had done away with the West Saxon Schriftsprache imposed by the Germanic elite in Anglo-Saxon times that some features of Brythonic Celtic in the spoken language, like the preference for internal possession, were able to surface.

The book is organised in nine chapters and contains three appendixes, a reference list and an index of authors and terms. Chapter 1 sets the scene for the ensuing analysis by introducing the basic theoretical aspects of the syntax (and semantics) of external possession, some typological considerations and a brief discussion of previous studies of external possession in early English. Chapter 2 is an overview of the uses of the dative case in Old English, including some which are not the main focus of the study, such as datives in ditransitive constructions, in impersonal constructions, expressing existence, and ‘free’ datives. Chapter 3 is a preliminary to the analysis presented in chapters 4 to 7, introducing some methodological considerations about the corpus, the design of the database and data retrieval. Chapter 4 presents the results of the quantitative and qualitative analysis of both external and internal possession structures referred to body parts in Old English as possessa, making a clear distinction between direct arguments (objects and subjects) and objects of prepositions, and clearly separating prose from poetry texts. The analysis also makes an important distinction between subject possessa with lexical verbs and subject possessa in copular constructions. Chapter 5 is a most interesting interlude delving into the question of possible changes in the range and use of DEP even before the earliest recorded texts in English. To this purpose Allen provides us with a comparative analysis of the variation between DEP and IP structures in Old English and other (older) Germanic languages such as Gothic and Old Saxon, as representative of Common Germanic. Chapter 6 replicates the analysis in chapter 4 with words referring to the mind and spirit in Old English, and chapter 7 explores the later developments of the DEP construction in Early Middle English, making a special point of the question of the preservation of the dative case and the problems for linguistic analysis associated with its overt marking in both case-impoverished and non-case-impoverished texts. Chapter 8 presents a very detailed discussion of the different theories accounting for the loss of external possession structures in English, from both an internal and an external point of view, assessing them critically in the light of the factual evidence provided in chapters 4 to 7. Chapter 9 wraps up the volume summarising the findings in the previous sections. The book ends with a full list of references and three appendixes containing, respectively, detailed information about the corpus (appendix A), a list of lemmas used in the searches for the body and mind possessa, the affecting verbs and the prepositions used in the searches for PObjs (appendix B), and some comments on the problems the author faced when searching for DEPs and IPs in the electronic corpora (appendix C).

The study of the interaction of the fundamental variables analysed by Allen in accounting for the frequency of DEPs and IPs in her Old English corpus (grammatical category of the possessa, text type, date and semantics of the affecting verb) proves that the widely held assumption that DPS were very common in Old English is just an impression produced by a fairly unrestricted use of DPEs with PObj possessa across the corpus, a fact also noted in passing by Ahlgren (Reference Ahlgren1946: 125). In both poetry and prose texts, DEPs with PObj possessa clearly outnumber the instances of DEPs with direct argument possessa in Allen's corpus, while there is also a clear increase in the frequency of IPs for either type of argument within the period. In this respect, the quantitative data in Allen's work also prove that – contrary to the standard assumption – it was IPs and not DEPs that were the default resource to mark an affected possessor in the Old English period, except probably with some specific verbs. The study points thus in the direction of DEPs of direct arguments as becoming increasingly marked during the Old English period and their association with individual verbs and idiomatic phrases, to the extent that DEP structures eventually became lexically governed by the late Old English period. Regarding the semantics of the verbs involved in external possession structures in Old English, Allen's analysis demonstrates that while DEPs of object body possessa were in general adversely affected, the possessor of a body part subject could exceptionally be unaffected or even positively affected. With mind words, however, this restriction did not systematically apply, as the small number of DEP examples with mind possessa found in the corpus are not limited to adversely affected possessors. Rather, most of them refer to a beneficially affected possessor. In general, the quantitative data for the Old English period in Allen's study show that the preferred structure to indicate inalienable possession with affected possessa in that period was IP and contradict the hypothesis that the decline of the DEP construction took place abruptly in the Middle English period. In this respect, Allen's figures for the incidence of DEPs in the m1 period of the Helsinki Corpus do not at all provide a picture of sudden disappearance, but show that the most noticeable decline in the use of the DEP took place in the late Old English period. Some of the tendencies observed for DEPs in late Old English are likewise confirmed in the early section of her Middle English corpus: DEPs with direct arguments are increasingly restricted to highly affected possessors and found only with a small number of verbs, while the context which had been more fruitful for the use of DEPs in Old English (arguments of PObj possessa) undergoes a substantial decrease in Early Middle English, that being – according to the American theorist – probably the cause of the popular perception of a sudden disappearance of DEP structures in Middle English. Allen's analysis of the m2 period of the Helsinki Corpus data (where hardly any examples of DEPs are found) confirms the view that the profound change in the grammar of speakers preventing them from generating DEP structures was accomplished between the m1 and m2 periods. This timing does not fit the conventional chronology for the loss of the dative/accusative distinction in English, which is conventionally placed at the onset of Middle English. That represents, in my opinion, one of the major breakthroughs of Allen's work, as it clearly dismantles the traditional explanation for the loss of dative external possessors in English as directly associated with the syncretism accusative/dative in the English language. A convincing alternative explanation is offered, however, as Allen notes the strict correlation that exists between the loss of DEPs and the general narrowing in the range of dative functions that took place in the m2 period (when the absence of DEP structures is most clearly recorded in her corpus). According to Allen, therefore, the loss of DEPs would be directly linked to the disappearance of bare NPs of different types marked in the dative case in functions other than indirect object, like (2) below (cited on p. 169), in which the NP in dative is part of the argument frame of the verb rather than an addition to it:

While many of these uses of the dative case were still found in Early Middle English (even in case-impoverished texts where the dative did not have a distinct mark, like god in (3a) below from the Ormulum), the tendency after about 1250 was to replace those bare complements in the dative by alternative constructions which made grammatical relationships clearer, like the prepositional phrase to god in (3b) (both examples cited by Allen on p. 171):

Therefore, Allen suggests that it is a general reduction in the range of structures using the dative case in English and not the loss of distinct case marking for accusative/dative that caused the DEP to seriously recede and finally to disappear.

Regarding the non-structural explanations that have been put forward to account for the advancement of internal possession structures to the detriment of external possession in the history of English, Allen provides us with a set of convincing arguments that Latin must not have played a significant role in it, and that the role of Old Norse and Celtic must be considered contributory, not causal. Her systematic comparison of the expression of possession in Old English texts and their Latin originals (when we deal with translations) shows that the expression of possession in Old English worked highly independently of the Latin models, as often a DEP in the translation replaced a Latin IP (a move contrary to the expected analogical pull), and in a considerable number of examples the Old English translations show variation between the use of DEPs and IPs when the Latin originals lack overtly expressed possession altogether. Regarding the possible influence of Norse as the cause of the decline of DEP in English, Allen admits that the reinforcement of IPs over DEPs might be seen as a consequence of a typical contact situation when incomplete acquisition favoured simplification rather than transfer (Thomason and Kaufman Reference Thomason and Kaufman1988), forcing the English population after the Norse settlements to opt for a syntactic alternative to indicate possession (IP) that both communities of speakers – English and Scandinavian – shared, to the detriment of DEP. However, the results of Allen's study indicate that contact with Scandinavians probably just accelerated a process that had started long before. Two fundamental arguments in this direction are the clearly observable decline in the incidence of DEPs in West Saxon, the Old English dialect least influenced by contact with Norse, and the fact that highly Scandinavised medieval texts such as the Ormulum or the Peterborough Chronicle display a considerable number of examples of DEP structures. Finally, considering the possible influence of Celtic on the syntax of the Anglo-Saxons after the Germanic invasions, Allen discusses in some detail in section 8.3.3 some of the most important tenets of the Celtic Hypothesis (especially the assumption that Old English texts did not faithfully represent the spoken language of the population), and presents alternatives to Vennemann's (Reference Vennemann, Lucas and Lucas2002) assumption that the loss of EPs in English was a consequence of the ‘Celtic syntax’ acquired by the lower strata of Anglo-Saxon society, which was only able to surface in writing after the Norman Conquest. Proponents of the Celtic Hypothesis put this in parallel with other syntactic developments in English like the rise of periphrastic do, unique within the Germanic family of languages but also a feature of Celtic, which only turned up in writing in the Middle English period. Allen points out in her study, however, a very important difference between inalienable possession structures and periphrastic do: if the latter was alien to the grammar the Anglo-Saxons brought to England but was acquired by the Celts who learned English, it was a likely candidate to be stigmatised as a Celticism, but if IPs expressing affectedness were a characteristic feature of Germanic languages, as attested by their use in very traditional and prestigious Germanic genres such as alliterative poetry, then there is no reason why IPs should be stigmatised at all. Allen also insightfully comments that if the hypothesis that IPs with affected possessors only boomed after the West Saxon Sprachschrift was obliterated (as the defenders of the Celtic Hypothesis maintain) were true, then there would be no satisfactory explanation for the fact that Ælfric (a conscientious writer who revised his works very thoroughly) clearly favoured the use of IPs expressing inalienable possession in his works. In sum, Allen's position with regards to the influence of Celtic on the expression of inalienable possession in English is that while contact with Celtic speakers may probably explain a certain increase in the use of IPs in Old English, it cannot be considered the triggering factor of an abrupt shift to a system allowing only genitives to mark inalienable possession in English.

Allen's text is pioneering in many respects: it provides the first solid empirical basis to test the different theories that have been put forward about the demise of the DEP in English; it convincingly links empirical findings with relevant formal aspects underpinning the construction; and it presents us with illuminating reflections on the moot points of the different internal and external explanations that have been produced to date to explain the decline of external possession structures in English. Plus, Allen's text furnishes the reader with a wealth of quantitative and qualitative data to prove her theories (including comparative analyses on the characteristics of DEP structures in a remarkable number of European languages), which will be of great interest to a wide readership, including language typologists, diachronic linguists and historians of the English language. I do not doubt that this volume will be the reference text for anyone interested in dative external possession structures in English for many decades.