by Keith Tse (Chinese University of Hong Kong)
Prepositional infinitives are an important type of clausal complementation in all Romance languages, especially the use of de-infinitive and ad-infinitive which are pan-Romance in their uses as non-finite clausal complements (Harris 1978:197-198, Vincent 1988:68-70, Ledgeway 2012a:179, cf. Meyer-Lübke 1900:426ff.). However, although Romance prepositional infinitives are widely attested across time and space, their Latin (or proto-Romance) origins are as yet unknown, since prepositional infinitives do not exist in Latin, apart from some very late and dubious examples which cannot be taken for granted (Diez 1876:201-202, Beardsley 1921:97). Nonetheless, there have been recent attempts to reconstruct proto-Romance prepositional infinitives, which are structurally equivalent to Latin prepositional gerunds/gerundives as suppletive markers of the oblique functions of the infinitive and the latter may be taken as precursors of the former (Schulte 2007:87ff).
In this post, I outline a proposal concerning the Latin origins for Romance prepositional infinitives whose diachronic formation displays striking parallels with and divergences from the famous English to-infinitive (Los 2005), a comparison of which raises new questions not only for non-finite complementation but also for mechanisms of syntactic change.
Prepositional complementation in Romance
The two most common types of prepositional complementisers in Romance are de-infinitives and ad-infinitives, which show different distributions; the former is used with all types of verbs, while the latter is restricted mainly to verbs that imply purpose and futurity (Meyer-Lübke 1900:426ff, 435ff; Beardsley 1921:97-99, 106-108, 150-151; Vincent 1988:68; 1999:7). This is illustrated in the following examples from Medieval Romance where de-infinitives are used with verbs of communication (verba declarandi), command (verba praecipiendi) and as prolative infinitives (verba prolativa), whereas ad-infinitives are only attested with the latter two (prepositional complementiser in bold):
1a) deneg-o de enuia-r-les ayuda
deny-PRET.3SG DE send-INF-PRO.3PL aid
‘… he denied that he sent them help.’ (La Primera Crónica General 679a33)
1b) confess-a d’ aver-lo fa-tto
confess-PRES.3SG DE have-PRO do-PERF.PTCP
‘he confesses that he has done it…’ (Rettorica p. 108)
1c) qui se dout-e d’ estre blasmee
REL.PRO REFL.PRO fear-PRES.3SG DE be.INF blame.PERF.PTCP
‘… who fears that he is being blamed.’ (La clef d’amors 2584)
2a) siempre contiend-e de val-er a cuitad-os
always strive-PRES.3SG DE protect-INF AD victim-PL
‘he always strives to protect the victims.‘ (La Estoria de Sennor Sant Millan 623)
2b) procaccia-ndo di riconcili-ar-si co-l Papa
strive-GERUND DE reconcile-INF-REFL.PRO with-DEF.ART Pope
‘striving to reconcile with the Pope.’ (Cronica fiorentina, p. 104)
2c) desirroit a vivre d-u sien
want-COND.3SG AD live.INF DE-ART.MASC.SG his.MASC.SG
‘… he would like to live with his.’ (Les miracles de saint Louis de Guillaume de St Pathus 5554)
3a) ell-os ordena-uan de pon-er
PRO-3PL order-IMPERF.3PL DE place-INF
‘… they ordered to place them.’ (La Primera Crónica General 87a47)
3b) pora esforç-ar a defend-er-se
in.order.to force-INF AD defend-INF-REFL.PRO
‘in order to force them to defend themselves.’ (La Primera Crónica General560b31)
3c) ordin-arono di fa-r-gli fa-re incontinente…
order-PRET.3PL DE make-INF-PRO make-INF incontinent
‘… they ordained him to be made to make him incontinent’ (Compagnia di S. M. del Carmine, p. 66)
3d) era-no costr-ett-i … a tagli-are selv-e
be.IMPERF-3PL force-PERF.PTCP-NOM.PL AD cut-INF forest-PL
‘… they were forced… to cut forests…’ (Vegezio 2, cap. 24)
3e) il fust contrei-nz a renoi-er la foy Jhesu Crist
PRO be.PRET.3SG force-PAST.PTCP AD reject-INF DEF.ART faith Jesus Christ
‘… he was forced to reject his faith in Jesus Christ.’ (L’histoire de Barlaam et Josaphat 1.1.46)
The main difference between de and ad, therefore, is that de marks both realis and irrealis clausal complements, whereas ad only marks irrealis complements, which may be projected back to proto-Romance. In the next section, I look at some Latin attestations which bear striking similarities to these Romance examples and may be taken as their precursors.
Prepositional complementation in Latin
Both Latin de ‘about, regarding’ and ad ‘to, towards’ are lexical prepositions; there are numerous examples from pre-classical and classical times where prepositional gerunds/gerundives are construed directly with verbs which are compatible with their lexical meanings of these prepositions (Johndal 2012). In the case of de, it denotes the content of propositions and is attested with numerous types of verbs that express indirect statements (prepositions in bold):
In this category, these are examples of verbs of saying and thinking (dicendi et putandi) that take de-gerund/gerundive expressing the content of the proposition, which can be reanalysed as indirect statements:
4a) primum tibi de nostr-o amico
first PRO.2SG.DAT DE our-ABL.SG.MASC friend-ABL.SG.MASC
placa-nd-o aut etiam plane
appease-GERUNDIVE-ABL.SG.MASC or even altogether
‘First I promise you about appeasing or even restoring our friend altogether.’ > ‘I promise you that I shall appease or even restore our friend’ (Cicero ad Atticum 1.10.2)
4b) qui de virgine capienda
REL.PRO.MASC.NOM.PL DE girl-ABL.SG capture-GERUNDIVE-ABL.SG
‘who wrote about capturing the girl’ > ‘who wrote that they would capture the girl’ (Gellius Noctes Atticae 1.12)
4c) tu de alter-o consulat-u
PRO.2SG.NOM DE another-MASC.ABL.SG consulship-MASC.ABL.SG
gere-nd-o te dice-re-s cogit-are
run-GERUNDIVE-MASC.ABL.SG PRO.2SG.ACC say-IMPERF.SUBJ-2SG consider-INF
‘you said that you were considering about running another consulship’ > ‘you said that you were considering running another consulship.’ (Cicero In Vatinium 11)
4d) nam vell-e se cum eo conloqu-i
for want-INF REFL.PRO with PRO.3SG-ABL converse-INF
de parti-end-o regn-o
DE divide-GERUNDIVE-ABL.SG kingdom-ABL.SG
‘for he wanted to converse with him (something) about dividing the kingdom.’ > ‘for he wanted to say to him that he would divide the kingdom.’ (Nepos Dion 2)
De-gerund/gerundive and ad-gerund/gerundive are used with certain verbs expressing the content of intention/purpose of the matrix subject:
5a) nos… labor-amus de aufere-nd-o mal-o
we work-PRES.1PL DE eliminate-GERUNDIVE-ABL.SG evil-ABL.SG
‘we strive about removing the evil…’ > ‘we strive to remove the evil.’ (Tertullian Adversus Hermogenem 11.3)
5b) ego enim te arbitr-or… statim esse
PRO.1SG for PRO.2SG think-PRES.1SG at.once be.INF
ad Sicyon-em oppurgn-and-um profe-ct-um
AD Sicyon-ACC attack-GERUNDIVE-ACC set.out-PERF-ACC.SG
‘for I think that you immediately set off in order to attack Sicyon’ > ‘for I think that you immediately set off to attack Sicyon’ (Cicero ad Atticum 1.13)
Verbs denoting command can take both de-gerund/gerundive and ad-gerund/gerundive in expressing the content and purpose of the command respectively, which may be reanalyzed as indirect commands (Panchón 2003:384-387):
6a) cum de muta-nd-o praecip-ere-t homin-e
since DE change-GERUNDIVE-M.ABL.SG order-IMPERF.SUBJ.3SG man-M.ABL.SG
‘since he ordered about changing the man’ > ‘since he ordered to change the man.’ (Augustine Sermones 9.8)
6b) ut consul-es populum cohort-are-ntur
so.that consul-NOM.PL people-ACC.SG encourage-IMPERF.SUBJ-3PL
ad rogation-em accipiendam
AD plea-ACC.SG accept-GERUNDIVE-ACC.SG
‘so that the consuls might encourage the people so as to accept the plea’ > ‘so that the consuls might encourage the people to accept the plea’ (Cicero ad Atticum 1.14)
6c) ad resistitue-nd-um non compell-it
AD re-establish-GERUND-ACC.SG NEG force-PRES.3SG
‘he does not force you so that you might re-establish it.’ > ‘he does not force you to re-establish it.’ (Augustine Epistulae 153.21)
The distribution of Romance prepositional infinitives hence seems to conform to Latin prepositional gerunds/gerundives where de in being the marker of theme/content is semantically more general and hence compatible with a wider range of verbs whereas ad as a marker of purpose/intention is only used with verbs that express command and purpose. These developments are strikingly similar to English to-infinitives, especially from a formal perspective, as discussed in the next section.
Prepositional phrases > prepositional infinitives
English to-infinitives are the prototypical example of non-finite complementation and it is widely held that to-infinitives are reanalysed in Old English (OE) from being purposive adjuncts to clausal complements (cf Latin ad-gerund/gerundive), which are particularly frequent with verbs of purpose and command (Los (2005:chapter 3)):
7a) tiligen we us to gescild-enne and us to gewarnig-enne
strive we us TO shield-DAT and us to guard-DAT
‘we should try to shield ourselves and guard ourselves…’ (HomS 44,158)
7b) on hwilcum godum tihst pu us to gelyf-enne ?
in which gods urgest thou us to believe-DAT
‘which gods do you urge us to believe in?’ (AELS (George) 148)
Furthermore, both Latin/Romance and English prepositional infinitives are the results of morphophonological erosion in the nominal paradigm, since the Germanic dative ending –enne following OE to is argued to be obsolete in OE (Los 2005:3-5) and the Romance infinitive, in contrast to Latin gerund/gerundive, likewise does not inflect for morphological case. In both cases, the nominal properties of the clausal complement are practically eliminated, which severely weakens the agreement between the preposition and its nominal complement (Roberts and Roussou 2003:105), which leads to their reanalysis as non-finite clauses. Furthermore, Latin/Romance de-infinitives represent a new pathway of syntactic change since, in contrast to English to-infinitives and Latin/Romance ad-infinitives, Latin/Romance de does not express purpose but is more semantically general in expressing the content of propositions, which not only yields its wider distribution in Romance but also reveals two distinct types of non-finite complementisers, one more purpose-oriented (to/ad), the other more neutral (de). Since non-finite complementisers are traditionally held to be low in the cartography of C-elements (Rizzi 1997), it may be argued that there are two functional projections in the non-finite domain (Mrealis/Mirrealis), which parallels the dual complementiser system in Romance finite complementation (Ledgeway 2012b). The Latin/Romance evidence, therefore, reveals a more sophisticated C-system, especially in the non-finite domain.
The use of Latin prepositional gerund/gerundive represents a new topic in Latin/Romance historical syntax which opens up many new avenues to the formation of Romance non-finite complementation, since although prepositional infinitives, which are plentiful in Romance, are not attested in Latin, their historical structural equivalents, namely Latin the prepositional gerund/gerundive, are widely attested in examples where they are re-analysable as clausal complements. It is therefore possible to account for the pan-Romance distribution of prepositional infinitives by expanding our search and analysis to Latin prepositional gerunds/gerundives.
Beardsley, Winfred, A., 1921, Infinitive Constructions in Old Spanish, New York, Columbia University Press.
Diez, Frédéric, 1876, Grammaire des Langues Romanes, vol III. 3rd ed, Paris, Libraire-Éditeur.
Harris, Martin, 1978, The evolution of French syntax: a comparative approach, London, Longman.
Johndal, M. (2012): Non-finiteness in Latin. DPhil dissertation, University of Cambridge.
Ledgeway, A. (2012a): From Latin to Romance: Morphosyntactic Typolog and Change. Oxford: Oxford University Press.
Ledgeway, A. N. (2012b): ‘La sopravvivenza del Sistema dei doppi complementatori nei dialetti meridionali’, in Del Puente, P. (ed): Atti del II Convegno internazionale di dialettologia-Progetto A.L.Ba. Rionero in Vulture: Calice, pp. 151-176.
Los, Bettelou, 2005, The Rise of the To-infinitive, Oxford, Oxford University Press.
Meyer-Lübke, Wilhelm, 1900, Grammaire des Langues Romanes. Tome Troisième: Syntaxe, Paris, H. Welter.
Panchón, Federico, 2003, ‘Les complétives en ut’. In: Bodelot, Colette, 2003, Grammaire Fondamentale du Latin. Tome X: Les propositions complétives en latin, Louvain/Paris/Dudley, Peeters: 335-481.
Reenan, Pieter, van. / Schøsler, Lene, 1993, ‘Les indices d’infinitif complément d’objet en ancien français’. In: Lorenzo, Ramón (ed), Actas do XIX Congreso Internacional de Lingüística e Filoloxía Románicas, Vol V, La Coruña: 523-545.
Rizzi, L. (1997): ‘The fine structure of the left periphery’, in Haegeman, L. Elements of Grammar, Netherlands: Kluwer Academic Publishers, pp. 281-337.
Roberts, I. and Roussou, A. (2003): Syntactic Change. A Minimalist approach to grammaticalization. Cambridge: Cambridge University Press.
Schulte, Kim, 2007, Prepositional infinitives in Romance: a usage-based approach to syntactic change, Oxford, Peter Lang.
Vincent, Nigel, 1988, ‘Latin’. In: Vincent, Nigel / Harris, Martin (eds), The Romance Languages, London, Croom Helm: 26-78.
Vincent, Nigel, 1999, ‘Non-finite complementation in Latin and Romance’, Paper presented at the Indo-European Seminar, Department of Classics, University of Cambridge, October 1999.