[Authors and titles are listed at the end of the review]
For me, as for many others, collocations are combinations of words with a high ʻmutual information scoreʼ: for example, and at the is a very frequent combination, but not a collocation—the combination is frequent because it consists of three very frequent words; whereas on the other hand, while being considerably less frequent in absolute terms, counts as a collocation, because on the other allows us to predict, with reasonable accuracy, that the next word will be hand.][1] Collocations, then, consist of word groups with high combinatorial and transitional probabilities. This volume focuses on a subset of collocations, ʻsupport verb constructionsʼ, as in come to an agreement, roughly equivalent to agree, where the verb come has a weakened, metaphorical meaning and marks the event as inchoative. Thus, though I had expected a broader range of topics to be covered, the book is nonetheless wide-ranging, thorough, and convincing in its arguments.
The introductory chapter, written by the editors, briefly summarizes and discusses the individual chapters. Next, Roland Hoffmann examines Latin support verb constructions from a typological perspective. For Hoffmann, the type animum aduertere does not count as a prototypical support verb construction because the noun is not a verbal abstract, which he views as a crucial factor. Support verb constructions have not yet been given much attention in language typology, so Hoffmannʼs undertaking is a valuable first step, even though the database is very small.
Cristina Tur looks at manus ʻhandʼ in various collocations For each combination of a noun with a verb we need to ask three questions: (a) can we compute the meaning of the combination compositionally, as a sum of its parts?, (b) if not, how transparent is the meaning?, and (c) how fixed is the formation? With these criteria in mind, we can distinguish between free combinations (statuam facere ʻmake a statueʼ); collocations (lexical: oua facere ʻlay eggsʼ; functional: bellum facere ʻwage warʼ; support verb constructions: bellum concitare ʻprovoke warʼ); phraseological units (turns of phrase: manus dare ʻsurrenderʼ; formulae: di te perdant ʻmay the gods ruin youʼ); and proverbs (summum ius summa iniuria ʻthe greatest right is the greatest injuryʼ).
Irene De Felice and Chiara Fedriani turn to ʻup/downʼ metaphors in Latin, especially with nouns involving anger, fear, love, hate, joy, and sadness. These emotional concepts are often envisaged as physical objects, living beings, entities that increase or decrease, or places. For instance, an emotion seen as a physical object can be lifted up or put down, or it can rise or fall. In these last metaphors, there seems to be some confusion because ʻupʼ and ʻdownʼ can both be inchoative or terminative; but this confusion is only apparent: the emotion can rise (inchoative) or fall (terminative), but it can also fall onto someone (inchoative) or rise and move out of someone (terminative).
Guillermo Salas Jiménez is interested in the question which verbs combine with bellum, proelium, and pugna. The verbs can be agentive (ʻtakeʼ, ʻdriveʼ, ʻenterʼ, ʻjoinʼ, ʻstartʼ) or non-agentive (ʻriseʼ, ʻcatch fireʼ). However, the distribution is anything but random and has to do with the fact that bellum is more abstract, while proelium and pugna are less so. With bellum, ʻtakeʼ and ʻriseʼ predominate, while with proelium and pugna it is ʻjoinʼ and ʻenterʼ.
Eusebia Tarriño Ruiz focuses on phrases like gratiam inire in polite requests; they typically occur in the future tense or the subjunctive, as is appropriate for this kind of speech act. She argues that gratiam inire contains a singular noun because we are dealing with an abstract concept, whereas we say gratias agere with a plural noun because there is a concrete reference to words of thanks. Gratiam habere is not normally passivized, so that gratiam inire fills that slot, but the latter also exhibits ingressive aspect.
The contribution by José Miguel Baños and Dolores Jiménez López, on operari and facere in the Vulgate, is the highlight of the volume for me. They notice that operari is first used as a support verb in translation calques with nouns like iniquitas, abominatio, peccatum, impietas, nefas, scelus, and iniustitia. Hebrew has two main support verbs, the general ʻāśāh and its poetic equivalent pāʻal, and in Greek these are normally rendered as ποιέω and ἐργάζομαι, respectively. In the Vulgate, they roughly correspond to facere and operari, but the latter is more common than one might expect from simply equating the two verbs with their Greek counterparts. Jeromeʼs more common use of operari may stem from a desire for variation. Incidentally, phrases like ἀνομίαν ποιέω are highly marked in classical Greek, where one would prefer a neuter adjective over the abstract noun; in the Septuagint, such phrases are common as the result of direct translation from Hebrew. In Latin, operari used to be an uncommon verb; there are barely 70 tokens before Tacitus and Gellius, and they are largely intransitive and poetic. By contrast, in the Vulgate, there are 269 tokens of operari, and they are now transitive. What started as a translationism spread to Christian literature more broadly, and it is remarkable that similar collocations still exist in (western) Romance in contexts with a Christian ethos, for example Spanish obrar mal or obrar verdad.
Iván López Martín examines support verb constructions in the Historia Augusta, a problematic work transmitted in two parts and written, according to the manuscript tradition, by six authors; however, modern scholars have sometimes argued for fewer authors, and sometimes just a single one. López Martín presents the number of support verb constructions per 1,000 words for the traditional six authors: Aelius Spartianus has 16.80, Julius Capitolinus 16.42, Aelius Lampridius 12.01, Vulcatius Gallicanus 9.80, Trebellius Pollio 11.13, and Flavius Vopiscus 10.76; there is thus a clear statistical difference between the four authors of the first part, when taken together, and the two of the second part. What is more, the first part does not only have a greater number of collocations, but a great deal more variety among them, especially with habere and esse. Since support verb constructions are unlikely to be a conscious focus for stylistic variation, it seems that the Historia Augusta was written by at least two different authors, one for each half, and potentially by more than that.
Tatiana Taous looks at four terms in the grammatical tradition, collocatio, idioma, periphrasis, and soloecismus. Idioma quickly became the special term for Biblical usages that are based on Hebrew constructions, while soloecismus refers to issues such as using the wrong case because two constructions are mixed up. For periphrasis, she quotes the charming definition by Sacerdos (gramm. 1.467.20): longiore oratione … quam necessitas postulat. Taous also studies the support verb constructions in various Colloquia and notes that often Greek support verb constructions correspond to Latin simple verbs and vice versa. She concludes that the writers probably simply picked the best-attested, most frequent constructions.
Alfonso Vives Cuesta and Lucía Madrigal Acero look at εὐχὴν ποιέω in post-classical Greek. In early and middle Byzantine texts, the active is far more common than the middle; but in texts that rewrite such models, active and middle are roughly equally frequent, and in late Byzantine texts, only the middle is used. In classical Greek, ποιέω is almost always used in the middle if it functions as a support verb. This allows us to distinguish πόλεμον ποιεῖσθαι ʻto wage warʼ from πόλεμον ποιεῖν ʻto cause warʼ. When later, more refined authors rewrite earlier texts, they turn the active in koine support verb constructions into the higher-style middle.
María Isabel Jiménez Martínez and Cantal Melis are interested in the most frequent support verbs in the history of western Romance, that is, facere, dare, and mittere.[2] In the 11th to 14th centuries, the picture is still similar across languages: in Spanish, hacer predominates with placer, dar with alegría, meter with miedo; and the same situation holds in Portuguese, where of course the forms look slightly different (Spanish hacer corresponds to Portuguese fazer etc.) and in Italian, where fare predominates with piacere but also with vergogna; dare predominates with gioia but is not uncommon with piacere or terrore either; and mettere is common with paura and terrore. All in all, then, hacer / fazer / fare is the most common verb, especially in Portuguese, but absolute numbers do not tell us the whole story because there are clear collocations, and these are not equally frequent across text types. In the 16th and 17th centuries, the modern situation begins to crystallize: in Spanish, hacer remains common with placer, but dar can be used here as well and expands its overall sphere drastically (e.g. to ʻfearʼ), with the other verbs decreasing; in Portuguese, we find a more complex picture, insofar as meter is still the normal collocation with medo, but no verb predominates overall; and in Italian, fare expands its usage dramatically.
Begoña Sanromán Vilas focuses on collocations in the history of Spanish. She argues that meter survives as a support verb mainly because of fixed collocations that continue, such as meter miedo. The verbs causar, producir, and provocar came into use as support verbs only later, in higher registers, where they survive, not least because they are semantically compatible with a wide range of nouns and do not have specific collocational preferences. The author also asks important questions that can only be answered in future research; for example, to what extent can we predict which support verb will be chosen for which noun or group of nouns, and what are the implications for language acquisition, be it acquisition of a first or a second language?
Not every chapter in this book will be of interest to every classicist, but there is something in there for everyone. I have learned a great deal from this volume and will recommend it to my more advanced students. It is well produced, with few typos and few issues with the Latin and Greek examples.[3] Some of the contributions in English are not always idiomatic, and given that this volume is multilingual from the outset, with contributions in Spanish, Italian, French, and English, I would have preferred that these authors write in their native tongues (I find it hard to imagine that there could be classicists who are unable to read German or Spanish, and I for one prefer beautiful Spanish over less-than-perfect English). But these minor concerns should not detract from what is, to all intents and purposes, a wonderful achievement.
Authors and Titles
- José Miguel Baños, Dolores Jiménez López, Isabel Jiménez Martínez, and Cristina Tur, ʻIntroductionʼ
Part I: Collocations, Typology and Cognitive Linguistics
- Roland Hoffmann, ʻLatin Support Verb Constructions: A View from Language Typologyʼ
- Cristina Tur, ʻCombinatoria léxica y lenguaje figurado: algunas consideraciones sobre colocaciones latinas en la frontera con la fraseologíaʼ
- Irene de Felice and Chiara Fedriani, ʻCollocazioni verticali: metafore di orientamento up / down nella lingua latinaʼ
- Guillermo Salas Jiménez, ʻColocaciones incoativas en latín y metáforas conceptualesʼ
Part II: Collocations and Applied Linguistics
- Eusebia Tarriño Ruiz, ʻColocaciones verbales y pragmática: gratiam inire en la petición cortésʼ
- José Miguel Baños and Dolores Jiménez López, ʻTranslation as a mechanism for the creation of collocations (II): the alternation operor/facio in the Vulgateʼ
- Iván López Martín, ʻThe use of verb-noun collocations as a criterion for attribution of authorship: the Historia Augustaʼ
- Tatiana Taous, ʻCollocations et constructions à verbe support chez les grammairiens latins et dans quelques manuels antiques: un impensé didactique?ʼ
Part III: A Diachronic View
- Alfonso Vives Cuesta and Lucía Madrigal Acero, ʻSupport verb constructions in post-classical Greek and sociolinguistics: a diachronic study of εὐχὴν ποιέω as a level-of-speech markerʼ
- María Isabel Jiménez Martínez and Cantal Melis, ʻContinuidad y cambio en las colocaciones del latín a las lenguas romancesʼ
- Begoña Sanromán Vilas, ʻLas colocaciones verbales en la historia del españolʼ
Notes
[1] I gratefully acknowledge that my research is supported by a Major Research Fellowship from the Leverhulme Trust (MRF-2022-031).
[2] Mittere undergoes a dramatic change, since in Latin mittere metum means ʻstop oneʼs fearʼ, while the Romance continuations have come to mean ʻcause fearʼ.
[3] On p. 71 (ex. 16b), fucum factum is treated as passive, but factum is a supine. On p. 113 (ex. 59), the preposition must be ab rather than ad. On p. 139, ex. 9a is from Plautus, not Seneca. On p. 224 (ex. 38c), operamini should be operemini.