Nominalization as a typological phenomenon: A comparison between Latin and Australian languages: Types and tokens

5.1 From derivational to inflectional morphology

Considering that the type of Latin supine ending in -(t)um meets us at the beginning of the literary tradition in clearly defined form and complete grammatical potency, used not only for the statement of a spatial goal but also to express the notion of finality and with verbal government of a dependent object, it would appear that the modulation of the nominal -tu-stems as an original derivational category into the function of inflectional morphology is something that has already occurred at some time in the past. In terms of our first hypothesis, however, the prerequisite for this categorial change is full productivity of the relevant pattern of word formation, that is, of action nouns ending in -tu-. Therefore for action nouns the implication of this assumption would however be, that with the completed categorial change the productivity of the type of word formation as such would wane only then to expand once more when the newly created inflectional form has become obsolete. The Latin supine is in so far a predestined subject for testing this hypothesis, since the entire body of philological data of the “Königsberger Programmes” of the 1850s is available in Richter’s collection; the evidence is as a whole open to survey; and, in the course of the history of the Latin language, an individual biography of this grammatical category can be sketched out.

5.2 Old Latin

Let us begin with a look at the situation of Old Latin: copious material evidence can be furnished, especially from the comic dramatists (Plautus, Terentius, Caecilius) but also from Cato, Ennius, Lucilius and the tragedians (Accius, Pacuvius): there is no indication of any restriction within the condition of its use, i.e. of its dependence on a verb of motion.

These examples can be represented as a word-field having the neutral archilexeme eo in its centre, surrounded by verbs of destination (advenio, advento) and inchoative verbs (proficiscor) as well as by sundry examples with special semantic marking: (com)migro, pergo, curro, profugio, prosilio.

Also among the causative (factitive) verbs of motion we come across so-called “neutral” examples such as do, veho, duco, fero, besides cases which, as it were, form the periphery of the lexical field and lead to other semantic regions: (pro)voco, accerso and, as an extreme case, polliceor foras (catillatum) (cf. Plaut. Cas. 551–552: operam uxoris polliceor foras quasi catillatum ‘I’ve promised the aid of my wife out of doors, as though to go lick dishes’). Some expressions occur so numerously that we may assign them a phrasal, idiomatic character.

For example:

We have now tested which and how many of the -tu-abstracts present in Archaic Latin as supines ending in -tum, are in use as action nouns in Latin literature. The result, in my opinion, is instructive: all of the following examples do not qualify because they were apparently blocked in their formation and their use as independent substantives by the concurrent supine-forms.

That, among the substitute formations given in brackets, a particularly large number of action nouns in -tio may be found renders the well-known distinction between abstract nouns ending in -ti- and -tu-, drawn by Benveniste (1948), questionable at least in the case of Latin. As is well known, the author, through detailed investigations of mainly Indo-Iranian, Greek, and Latin word-material established a categorial semantics for the -ti- forms, “realisation objective” or “fait objectif de I’accomplissement,” and correlated it with the agent nouns ending in -tōr (Latin -tŏr) as designation of the “auteur d’un acte.” At the same time, in the tu-derivatives he perceived a general meaning, “disposition/capacité subjective,” which predisposed them to the function of the Latin supine and corresponded on the side of the agent noun with the -tēr- suffix as expression of the “agent voué a une fonction.” Where the -tu- forms stay on in use, they are mainly lexicalized examples, which have semantically shifted away from the predictable meaning of a deverbal abstract noun.

5.3 Classical Latin

But let us go on to the next step, to the period of Classical Latin: here we get the impression of stagnation in the development of our category. The dominant authors of this period, Caesar and Cicero, show a striking economy in the use of the supine ending in -tum. Where it seems to occur with any frequency we are faced not with productive new formations but with the quotation of set phrases (idioms) such as cubitum ire (discedere), salutatum, sessum ire, admonitum venire in Cicero, and mittere rogatum, convenire gratulatum, frumentatum or pabulatum progredi in Caesar. More often, however, these authors make other constructions serve their purpose: they use ad or the postpositive causa with the gerund or, above all, the newly proliferating gerundive construction (e.g. ad ludos spectandos / ludorum spectandorum causa venire); they use the increasingly popular, typically Latin, future participle which, in effect, continues the temporal grammatical sense or ire (e.g. ludos spectaturus venio), and, finally, the master-key formula of a purpose clause with ut (venio, ut spectem ludos). Varro, with his interest in the old language, has more frequent use for the supine, while, in the case of Sallust, as Skard (1932) points out, there is a constant increase in the use of this constructional pattern to match the growth of his archaizing tendency. Remarkable in this context is the lavish use of the supine in the work of the historians Cornelius Nepos and Livius; the latter, particularly, reanimates this construction in all the variety it has for us in old Latin. This is perhaps explained by the stylistic criticism of antiquity which occasionally stigmatized the author’s partiality for constructions which ignored the classical ideal of concinnity, a propensity which also expressed itself in that plentitude of participial phrase characterized as lactea ubertas (‘milky richness’). A further feature of Livy’s style was Patavinitas, i.e. the provincial smack of his native Patavium (Padua), declared by C. Asinius Pollio (Quintilian 1. 5. 56; 8. 1. 3) to typify Livy’s writing. We have already referred to the restriction of the supine to a few standard phrases in the case of Caesar and Cicero and interpreted them as the symptom of a gradual fossilization and lost transparence of this syntactic category. Similarly explicable is, on nearer examination, yet another phenomenon observable in the work of a few authors in the literature of classical Latin. The stock-list of governing verbs of motion, originally clearly defined by their semantic features and even representable as word-fields, is occasionally given up. These cases, clearly bold exceptions from the norms of grammar, to my mind exemplify centrifugal tendencies and a growing loss of motivation for the constructional type. Consider, for example, two sentences from Sallust and Vergil, respectively:

In philological commentary the example from Sallust is explained by analogy with ultum eo; ire, in this case, being taken as implied. I think, however, that another explanation is possible. The archaizing Sallust makes of the contemporary equation of the supine with ut-clauses, with the regular substitution of the latter for the former, an occasion for the use of the supine even in cases where we would expect not an ut-clause expressing finality but, rather, a clause expressive of demand dependent on hortari. This procedure is best described as hyperarchaism.

The second example from Vergil’s Aeneid, 9, 240f.: ‘If you permit to use fortune to see Aeneas and to visit the castle of Pallas’ – so irritated even the Vergil exegetes of antiquity that they sought to save their philological sanity and their author’s honor, grammatically speaking, by assuming dependence of the supine quaesitum on a participle euntis found in verse 243.

The explanation that lies nearer to hand is of course the simple one of a substitution here too of the choice and elegant-seeming supine for an ordinary clause with ut or a construction with ad and the gerund or gerundive. Here again, in the process, the concomitant feature of finality has edged out and annulled the genuine motivation of a directional specification.

5.4 The (gradual) decline of the supine in Late(r) Latin and in Vulgar Latin

But what of the further course of the Latin language? In the works of the Augustan poets (Horace, Tibullus, Propertius, and Ovid) the supine only occurs sporadically, and the historians of the period of Silver Latin (Curtius, Tacitus, Suetonius, and Florus) too offer only very few examples. The agrarian writer Columella provides, unlike his literary predecessors and models, Cato and Varro, only one solitary example (9,5,1). The historian Velleius Paterculus is the first author from whose works the supine is completely absent. In Petronius there is one instance to be found in the narrative part of his Satyricon (71,8); the fact that no example appears in the Cena Trimalchionis shows how foreign this construction has already become in the spoken language.

The supine experienced a certain renaissance with Gellius and Apuleius in the second century; which is scarcely to be wondered at as these authors, declared archaists, not only hawked old Latin materials (for example: from Cato and the ancient dramatists), but also in their own style took on an antique coloring. Virtually no relevant data are offered by late Latin poetry: There, in addition to the substitute expressions mentioned above, a construction with the infinitive of causality becomes more common (hence, a new type, e.g. venio ludos spectare). Revealing is also the literature of the language of colloquy, produced by authors of Vulgar Latin. Of the Cena Trimalchionis we already have affirmed an absence of examples; in the translations of the Bible we only find a few single examples frozen into formulae: thus, there appears the ubiquitous cubitum ire in Gen. 19,4, as well as the equally idiomatic potum dare in Matth. 10,42, both in the Vulgate. In the Peregrinatio Egeriae 24,12 we find a reponent se dormito, an equally formulaic phrase, which, however, in its phonetic details (u > o, loss of final -m) has something of the color of contemporary use. Gregory of Tours (Franc. 5,15) fits in here with his ulto irent: moreover, he offers an example of a clarifying of the accusative of the supine form by the preposition ad in the phrase deducat equos ad pastum (Franc. 3,15); something which fits well in the picture of a gradually developing analytical inflection in Late Latin. At the same time, the form pastum, because of this, reenters the categorial role of verbal noun and thus becomes an element of derivational morphology once more. This process grows still more obvious in the phrase ad potionem veniet in the Mulomedicina Chironis (444 and 503). Here the deverbal abstract noun potio takes over the traditional function of the formulation ending in -tu.

5.5 Middle Latin and Romance languages

With this, however, the history of the supine in Latin comes to an end: in living use of Late Latin this type of construction has disappeared completely, and if it partially returns in Middle Latin, that is due to reminiscences of classical Latin and artificial attempts to revive it. Significantly, there is no trace of the supine in Romance languages – except Romanian, where it might be due to an inherited archaism of the Latin of the Balkan periphery.

With the quiet washing away of this grammatical category in (spoken) Late Latin, however, the -tu- formations functioning as nouns turn up again – and with this the second part of our hypothesis is fulfilled. Extensive investigations of the lexicon have shown that just at the time when the supine had shrunk to a few vestigial formulae, before its final total disappearance, a veritable wave of new (i.e. not previously attested) deverbal derivations relating to the suffix -tu- swells up. I have collected several examples of them under (7); there exists in every case in the old Latin word-stock a homonymous supine that, however, at this late time – as shown by our texts – was long out of use.

5.6 Reasons for loss of productivity

I have now given a lightning sketch of the Latin supine and found confirmed my hypothesis concerning the changing productivity of word formation types which come to be regarded as elements of inflectional morphology. The question next arises why this grammatical category after its heyday in old Latin suffered so swift a loss of popularity. Coming, in the end, to be given up altogether. Seeing that hardly any work has been done on this problem, it follows that my own meditations on it are of a predominantly tentative character. The fact that a variety of possible reasons may account for the brief life of the Latin supine seems to me no disadvantage. I am, in principle, no defender of grand single reasons to explain things, inclining, in fact, to the view that a cumulative causality more genuinely accounts for the phenomenon of linguistic change.

5.7 On the hybrid character of the supine

In the present case we may start out from the point of the never entirely settled, hybrid character of the supine, positioned between the categories of noun and verb. We have the evidence of ancient grammarians to support us here, an eloquent testimony being available in the history of the naming of the supine which is detailed for us in the admirable researches of Benveniste (1932: 136–137) and Wackernagel (1926: I,276ff.). The name, supine, derives from the Greek hyption, a metaphor taken from the language of the wrestling schools. It was used at first for verbs of the type audeō, gaudeō, or veneō, verbs which fluctuated between the active and passive voice, and was subsequently transferred to the supine proper, also neutral of voice.

Diomedes (G[rammatici] L[atini] I 342,4), in view of its hybrid nature, speaks of a participalis modus, for the category shares both nominal and verbal features, while Charisius (GL II 175,27) in relation to this entirely unclear situation speaks of adverbia qualitatis, including the gerund in this definition.

The nominal trait of the supine is explicitly stressed by the grammarian Priscian V,47f. (Gramm. Lat. II 412,3ff.), when he calls it an accusative without preposition which, like the names of cities, responds to questions of direction. As for the so-called supine II, he equated it with a verbal substantive: mirabilis visu for him means exactly the same as mirabilis visione.

Grammarians apart, other Latin authors too demonstrate to us over and over again, that they ascribe features of a noun to the supine. As one of the earliest of them we mention Cato, who in de agr. 5, 5 says of the vilicus: primus cubitu surgat, postremum cubitum eat. In this case, the ablative cubitū ‘should rise from the bed’ obviously corresponds to the supine cubitum, dependent on the verb of motion eat. Even a writer like Gellius, an archaist, as we know, and given to frequent use of the supine, clarifies it occasionally by introducing a preposition of direction, thus making obvious his conception of it as a noun: e.g. in pastum profecta est (2,29,13).

At least three factors – as far as I can see – militate against a durable validity for the supine conceived as noun.

The simple accusatives of direction responding to the enquiry “where to?” could at the time of the emergence of this category (partly too in literary old Latin) be formed also from appellatives/names. However, with the passage of time, certainly at least in the period of classical Latin, they are limited to place-names, particularly cities, and to a few adverbially frozen forms like domum ‘homeward’, rūs ‘to the country’, bellum ‘to war’. Forms like pastum ‘for grazing/to graze’, salutātum ‘in greeting/to greet’, or nuptum ‘for marrying/marriage’ lose in the course of this development both syntactic motivation and transparency.

5.8 The criterion of verbal construction

The dependence of predominantly accusative objects on supines as transitive verbal abstracts ending in -tu- (ex. [15]) was plausible at a time when other nouns of action showed verbal agreement as well (ex. [16]), therefore precisely in the period of Old Latin:

As constructions of this type are restricted to Old Latin (Late and Vulgar Latin texts allow partial revitalization of this pattern only with agent nouns), the supines as nouns, because also of their accusative objects, get shunted into a syntactic siding. In contrast, one may consider the supines ending in -tū, which have no verbal government (*Difficile est hoc factu; but: difficile est hoc facere): these maintain their position considerably longer in the history of Latin, showing a constant increase in use from Old Latin to Classical language.

5.9 The tendency toward categorial change of meaning

As I have suggested elsewhere, in various of my own studies on word formation, nouns of action generally show an inherent tendency toward categorial change of meaning, This development, for which the term “drift,” going back to Edward Sapir, seems convenient, tends to proceed through the level of resultative noun (nomen acti) and in many cases reaches the level of concrete noun (interpretable as instrumental or local), in certain cases achieving even an agentive reading. Thus we find vectis originally a noun ending in -ti-, meaning ‘motion forward, transport’ which becomes concrete as an instrumental noun meaning ‘lever, Jack’. The younger class of nouns ending in -(t)iōn, still productive in Latin, is not exempted from this general development: cp. mānsio the actional version of which ‘staying, stop’, becomes concrete as ‘place of residence’ in Late Latin, to end at last as the everyday word for ‘house’ in French maison. This development in Latin also did not spare abstract nouns: pastus not only means ‘(the) grazing’ but also ‘grazing ground’, cubitus also has the concrete sense of ‘bed, couch’, exercitus almost ceases to mean ‘exercise’ (except occasionally in the plural), and becomes, rather, ‘army’.

It is this fact, the existence of such highly lexicalized examples (see in our earlier list [3] gestus, habitus). which, in my opinion, provides a further argument against the expansion, or even the preservation of the category supine in later times.

5.10 Opinions and terms of Latin grammarians

We have already emphasized the hybrid character of the Latin supine as it oscillates between verb and noun, supporting our position with the opinions and descriptive terms of Latin grammarians. We shall now supplement these with quotations and points of fact which bring into stronger relief the verbal character of this category.

With this, he confirms the observation of modern philologists that oratum ire with its notion of future tense is interchangeable with the phrase oraturus est. This observation may be supplemented by the well-known fact that the so-called passive future infinitive in Latin, which, admittedly, its meagre occurrence considered, could never have been popular, has grown out of an infinitive formation based on the impersonal itur ‘one goes’ and a supine taken as a directional accusative.

That in the course of time a shift in the division into constituents took place, which made the directional case of the supine an element of the predicate phrase and the accusative object dependent on the supine the subject of an accusativus cum infinitivo is shown by the word order of both the old Latin examples that follow:

5.11 Instances of ambiguity between supine and perfect passive participle

If the quasi-subject role in these sentences were filled by a masculine or neutral noun or pronoun used in the singular, one might take the forms datum or deductum not as supines, but, more likely, as accusative participles of the perfect tense ending in -tus. And, here, we are on the trail of a further source of the gradual decline of the supine. The phonetic law dictating the raising of a short o in final syllables before m, s, r to an u, which inscriptions allow us to date in the third century BC, leads to the homophony of the supine formations with participial constructions. Thus supines (that is to say, certain morpho-syntactic and semantic conditions being fulfilled) could also be understood as participial constructions. We have thus the following ambiguous cases from Old Latin:

As Strunk (1962) has clearly demonstrated, such ambiguous cases become syntactic “switching points”; “Das Supinum wurde als eine adjektivische prädikative Ergänzung zum Objekt aufgefaßt…, Das neue Verständnis der syntaktischen Fügung führte nun zu einer völligen Adjektivierung des Supinums mit Genus- und Numerusattraktion” (Strunk 1962: 456).

This is demonstrated by the following three examples:

Again Strunk: “Diese scheinbar rein partizipialen Objektsprädikativa sind genau wie die prädikativen Gerundiva nach Verben wie dare, curare, usw. nichts anderes als adjektivierte infinitivartige Verbalsubstantiva” (Strunk 1962: 456).

5.12 Conclusions

With the ambiguous “hinge-forms” discussed above we have an additional causative factor for the continuous decline of the supine at a time when the phonetic system had initiated a development towards the homophony of this category with certain adjectival participle forms. We have reason therefore to assume that behind a further number of participial forms of the passive perfect, predicatively used, there lurks concealed a supine ending in -um, either disguised or misunderstood. A textual-critical examination of the relevant cases across the range of Latin literature would be, in truth, a formidable undertaking but one very rewarding in the circumstances.

7.1 Final remarks

Hopefully, this section shows that Australian languages exhibit different forms of nominalizations. Occasionally examples could be found where nominalized verbs assign a different case than the verb in a clause.

8.1 Semantic development as a drift phenomenon

A model that I developed some time ago for my own investigations of action nouns in a couple of languages – especially Modern German – provides a successive multi-stage development from an abstract meaning (as a process, state or factual circumstances) to a concrete notion following a scalar and irreversible procedure over mostly ideally six stages or steps:

(i) Nomen actionis → (ii) nomen acti/rei actae → (iii) concrete materialized result as a tangible object → (iv) instrument → (v) place maybe considered as an instrument with a local dimension → (vi) (group of) person(s)

This type of gradual change or rather mutation seems to be a true drift phenomenon in the acknowledged terminology of Sapir (1921: 168–170).

Drawing from contemporary German, this development can be illustrated step by step in the following chain: Bewilligung, Befreiung, Beseitigung – Rettung, Verwandlung, Besprechung – Zeichnung, Radierung, Verwundung – Heizung, Bekleidung, Ernährung – Wohnung, Behausung, Siedlung – Regierung, Versammlung; Bedienung, Begleitung.

To supplement this list with some comments, first I point at such types of pairs where merely the co-text, the context or the syntactic behavior makes a difference as far as the intention of the speaker/writer and the interpretation of the listener/reader is concerned. In a phrase die Rettung eines Menschen, the (result of the) actional notion becomes quite evident, whereas in the case of die Rettung wird bald kommen, the concrete institution with its equipment (car, crew, etc.) becomes more plausible. Both by the respective verb of motion (kommen) and the empirical pragmatic knowledge of the world. Likewise, the noun Erziehung can be understood both as a process and the result of this procedure. Cf. die Erziehung der Kinder ist eine schwere Aufgabe versus Diese Kinder haben eine gute Erziehung. In the case of Regierung with its threefold notion (‘governing’ – ‘government’ – ‘team of politicians; parliament’. A genitive attribute (Österreichs) might be interpreted either as a subjective or objective case. A similar situation appears in Russian, where sometimes the syntactic properties of a derivation enable the speaker/listener to differentiate between two possible readings. In the deverbal compound rukovodstvo when signifying “direction” as a procedure the nominalization is constructed with instrumental case just as in a syntagma involving a verb (rukovodstvo institutom), whereas when it means a (group of) person(s), adnominal genitive appears (rukovodstvo instituta). A similar doublet of constructions appears in vladenije: ‘mastery’ (jazykom ‘of language’ instrumental case) versus ‘possession’ (knig ‘of the books’ genitive plural).

For the genesis of such ambiguous, i.e. personal formations with personal interpretation or reading, sometimes a combination of syntactic and pragmatic reasons can be held responsible. In the case of Begleitung ‘company’ versus ‘accompanists’, a gap between the production of the utterance and its apperception/interpretation may be the (or at least a) reason for the ambivalent notion. In an expression Ich habe Hans mit neuer Begleitung gesehen the proper meaning of the deverbal noun is the process of being accompanied by a person. But since this “action” implies an “agent,” the nominalization can be understood as accompanist, which finally becomes lexicalized in this personal (individual or collective) meaning. For similar cases in the area of semantic change cf. Leumann (1927: 105–118).

In Latin, the scale of increasing concretization is attested for example in the semantic history of the item legatio: the semantic drift starts with ‘message’ (legationem suscipere), is continued as ‘task’ (legationem referre) and ‘office’ (legationem administrare) and ends up with (collective) agentive notion ‘ambassador, legation’ (legationes sociorum reverti iubentur). In another case this kind of successive shift appears in cenatio ‘feeding’ with its further readings ‘meal’ (concrete) and ‘dining room’ (location).

8.2 Some instances of lexicalization

Lexicalization in conventional and widely accepted linguistic terminology holds for cases when a lexeme though derived in a transparent way from a clear-cut base (deverbalization, denominalization, deadjectivalization) is not yet predictable as far as its special meaning is concerned. Whereas lectio from a semantic point of view is the result of the “intension” of legere together with the categorial meaning of action nouns, the items legio as a military term and lex ‘law’ (in a sense of a collection of orders, warnings, etc.) became opaque having drifted into the lexicon (both as a book and a mental object) and therefore must be learned as separate units. In most cases, lexicalization leads to a stricter and neater meaning (with a more specific intension, i.e. the internal content of a term or concept), whereas the extension (i.e. all the things or concepts to which it applies, also including the frequency of usage) decreases. The agent noun legulus is not the common word denoting a collector in its broad sense of application, but is specialized for a slave trained and assigned to collect olives fallen from the tree. Quite similarly factus (u-stem) exhibits another specialization of notion, since in a certain period of language history and within a sort of texts, it cannot denote the process of making/producing something in general but rather points at the pressing of olives to gain oil.