Participial perception verb complements in old English.

Lowrey, Brian

1. Introduction (1)

My intention here is to take a preliminary look at structures in Old English (henceforth OE) in which a present or "first" participle appears with an accusative NP in the complement of a verb of direct perception, such as (ge)seon or (ge)hieran, as in the following example:

(1) He pa gangende bi galilea sae geseah twegen gebroper ... settende nett in sae

'Walking by the sea of Galilee, he then saw two brothers ... casting a net into the sea'

(HC, Rushworth, 342-344)

I take the risk of using the term "Participial Perception Verb Complements," although as we shall see the relationship between the perception verb, the NP and the participle needs to be examined more closely. Indeed, examples such as (1) give rise to a number of questions.

First of all, we need to ask what kind of perception they describe. Perception verbs in OE, much like their Present-Day English (henceforth PDE) counterparts, occur with a wide range of complement types, and, as in PDE, different types of perception verb complement tend to encode different kinds of perception. For instance, the direct perception of entities, typically, is expressed both in PDE and in OE by an NP complement, whereas an infinitive complement of the type dubbed "VOSI" by Visser (1973) (2) (traditionally referred to as the "AcI" construction) is used to evoke the perception of an event. Since both structures tend to be used to relate direct sensory perception, they will henceforth be termed "Direct Perception Reports" or DPRs. They contrast in this respect with other types of complement, such as finite complements introduced by that (or to- infinitive complements in PDE), which, when used with perception verbs, evoke a kind of cognitive process rather than perception (3). I shall therefore consider these to be non-DPRs. Structures such as (1) in OE appear to be DPRs, but of what? Do they describe the perception of an event, in this instance the casting of a net by two brothers, or that of an entity, two brothers, who happened to be casting a net? This is perhaps the first question to be asked here.

Of course, this issue will have consequences for the syntactic analysis of constructions like (1). What exactly is the complement of the perception verb? Is it the NP alone, or some kind of combination of the NP and the participle? And what is the status of the participle? Is it a strictly verbal element, of which the accusative NP is the subject, or is it more adjectival in nature, modifying the NP? This is the question raised by Nickel (1966), and again by Mitchell (1985: [section][section]974-982) when he evokes "the problem of deciding whether the present participle is to be taken as adjectival or verbal".

We can also examine OE participial constructions in the light of what we know about apparently similar PDE structures. Declerck (1982) shows that there are in fact three different syntactic constructions that can give rise to the same surface string:

(1a) He saw two brothers casting a net into the sea

The first of these would be the aspectual equivalent of the VOSI infinitive form (He saw the event of two brothers casting a net). There also exists a second construction in which the participle is what Declerck (1982: 16-17) and Felser (1999: 68-71) call a "free adjunct", meaning something like: 'He saw two brothers, as they were casting a net.' Finally, Declerck identifies a third construction in which the participle acts as a kind of adnominal modifier, or "pseudo-modifier," closer in meaning to: 'He saw two brothers who were casting a net (4)'. These constructions have different syntactic and semantic properties in PDE5. To what extent do the same types of pattern occur in OE?

Perception verb constructions are interesting, too, from a pragmatic point of view. Kirsner and Thompson (1976) highlight the importance of pragmatic inferencing in the interpretation of perception structures in PDE, particularly where the perception of events is concerned. And we can observe a tendency in a number of European languages for constructions to be pressed into service as DPR markers although, superficially at least, this does not appear to be their original function. Take, for instance, the following examples from Russian and German:

(2) [TEXT NOT REPRODUCIBLE IN ASCII]

'We heard how he plays [We heard him playing] the guitar'

(3) Ich horte, wie das Flugzeug uber uns flog

'I heard how the plane flew [I heard the plane flying] over us'

Despite the presence of the words [TEXT NOT REPRODUCIBLE IN ASCII] and wie, corresponding approximately to English how, both (2) and (3) can be used to mean that what was seen or heard was not so much the manner in which the event took place, as the event itself. In other words, they have come to be grammaticalised, to a greater or larger extent, as DPR markers.

A further example is supplied by the so-called "relatives de perception" in French, as in (4) (to borrow a well-known example from Kirsner and Thompson 1976):

(4) Je l'ai entendu qui tuait le cochon

'I him-heard who was killing the pig'

(4) can be used to describe the same event as (4a), with an AcI construction

(4a) J'ai entendu le fermier tuer le cochon 'I heard the farmer kill[ing] the pig'

In (4), the clitic pronoun l', which refers to the farmer, is from a syntactic point of view the direct object of the verb entendre, 'hear'. Nonetheless, as shown by Radford (1975), and by Miller & Lowrey (2003), this construction can be used to express the perception not so much of the entity but rather of the event in which the entity is involved, as in (4), where what the speaker heard where most probably the cries of the pig rather than any sound made by the farmer. A construction originally expressing the perception of an entity [X], in the process of performing an action [Y], gives rise to the inference that the event [X do Y] has been perceived. I shall ask in this paper whether OE participial perception verb constructions (which for the sake of simplicity I shall henceforth refer to as "VOSende" by analogy with VOSI) give rise to a similar inference, and, if so, to what extent it has been lexicalised by a grammaticalised NP + present participle combination.

2. Evidence for the grammaticalisation of VOSende

It will be my contention here that VOSende had at least begun to undergo some form of grammaticalisation in OE times. In an attempt to show this, I shall look at two factors, in particular: evidence from what I shall call "parallel" occurrences with VOSI structures, and from co-ordination facts.

2.1. Evidence from parallel occurrences

I use this term to describe the manner in which VOSende is used in very much the same way as a VOSI construction, sometimes by the same author, within the same text, to describe either the same or a very similar event. Examples of this phenomenon are numerous in OE prose. Consider the following examples, from Aelfric's Homilies, in which the same verb, standan, appears on each occasion:

(5) Se halga Stephanus weard pa afylled mid pam Halgum Gaste, and beheold wid heofonas weard, and geseah Godes wuldor, and pone Haelend standende aet his Faeder swidran

'The holy Stephen was filled with the Holy Ghost, however, and looked towards heaven, and saw the glory of God, and the Saviour standing on the right of his Father'

(Aelfric, Homilies 46: 28)

(6) se forma martyr Stephanus cwaed, paet he gesawe heofonas opene, and done Haelend standan on his Faeder swidran

'the first martyr, Stephen, said that he saw the heavens open, and the Saviour stand[ing]6 on his Father's right'

(Aelfric, Homilies, 308: 35)

(7) and he cwaed, "Efne ic geseo heofenas opene, and mannes Sunu standende aet Godes swidran"

'and he said, "Behold, I see the heavens open, and the Son of man standing at the right hand of God"'

(Aelfric, Homilies, 46: 31)

(8) se halga cydere Stephanus cwaede paet he gesawe mannes bearn standan aet Godes swydran

'the holy martyr Stephen said that he saw the Son of man stand[ing] at God's right hand'

(Aelfric, Homilies, 48: 10)

Quirk & Wrenn (1957: 86) mention a tendency for the infinitive to be used "with verbs of motion, rest, and observation, often with durative aspect". Geseon, apparently, is included in the category of "observation" verbs, although there seem to be no aspectual considerations involved here. Each of the above examples offers a description of the same event, Stephen's vision of Christ standing at God's right hand. If we assume the VOSI structures of (6) and (8) to be DPRs of this event, then presumably the VOSende sequences in (5) and (7) are fulfilling the same function (7). The two constructions, in this context, appear to be functional equivalents.

If VOSI, in OE as in PDE, is the syntactic form associated with event perception, (6) and (8) tell the hearer directly that Stephen witnessed the event, whereas in (5) and (7) he arrives at the same conclusion via the inference that, when seeing the Saviour or the Son of Man, as he was standing on God's right, he has in fact witnessed the event of the Saviour standing on God's right.

Similar alternations can be observed in other texts, too. Here are some more examples from St. Mary of Egypt, all involving the same verb, standan:

(9) Pa he hine beseah, pa geseah he unmaettre micelnysse leon wia paere halgan lichaman standan

'When he looked at him, then he saw a lion of tremendous size stand[ing] by the holy corpse'

(St. Mary, 116: 920)

(10) Da geseah ic of paere stowe pe ic on stod paere halgan Godes cennestran anlicnysse standende

'Then, I saw from the place where I was standing a likeness of the holy mother of God standing'

(St. Mary, 90: 490)

(11) Pa geseah ic tyn geonge men aetgaedre standende be pam warude 'Then I saw ten young men standing together by the shore'

(St. Mary, 84: 415)

As a further indication that the two constructions did indeed function as functional equivalents, in the surviving Latin text closest to that from which the OE St. Mary was translated (according to Magennis, 2002: 30-35) the same NP + present participle construction corresponds to (9), (10) and (11).

Nor should it be thought that the alternation was limited to the sole verb standan:

(12) Pas engla werod beod aeteowde gesewenlice urum gesihdum on to-cyme oaes strecan Deman, ... Ponne we geseoo mannes Bearn cumende on wolcnum, mid micelre mihte and maegenorymme

'These hosts of angels will appear visible to our sights at the advent of the severe Judge, ... Then we shall see the Son of man coming in clouds, with great might and majesty'

(Aelfric, Homilies, 610: 23)

(13) Se hyra, seoe nis riht hyrde, he gesihd pone wulf cuman, and he forlaet oa scep and flyho; and se wulf sum gelaeco and oa oore tostenco

'The hireling, who is not the true shepherd, sees the wolf com[ing], and he abandons the sheep and flees; and the wolf seizes one, and scatters the others'

(Aelfric, Homilies, 238: 14)

(14) ... pa geseah ic on sumere tide micele meniu affricana 7 egypta togaedre yrnende swa swa to sae

'then I saw on one occasion a great multitude of Africans and Egyptians running as if to the sea

(St. Mary, 84: 387)

(15) Pa geseah ic soolice on aernemergen hi ealle anmodlice to paere cyrcan yrnan

'Then truly I saw them all, early in the morning, run[ning] eagerly to the church'

(St. Mary, 88: 454)

In the light of examples such as (5)-(8), it seems very probable that in these instances too VOSende is being used in much the same way as the VOSI construction to express the DPR of an event. The respective variational spaces (Smith 1996: 43-47) of the two constructions have begun to overlap.

2.2. Evidence from coordination facts

Alongside parallel occurrences such as those listed above, one also finds instances where VOSI and VOSende are coordinated with one another in the complement of a perception verb. A good example of this phenomenon is provided by (16):

(16) Ic geseah pa englas, pe eower gymdon, dreorige wepan, and da awyrigedan sceoccan blissigende on eowerum forwyrde

'I saw the angels who had charge of you weep[ing] drearily, and the accursed fiend rejoicing in your destruction'

(Aelfric, Homilies, 66: 35)

The speaker's purpose here is to contrast the reaction of the angels and that of the devil to the corruption of two brothers, Eugenio and Attico (the referents of eower), by stating what he saw the angels doing, and what he saw the devil doing. Once again, if the infinitive complement denotes the perception of an event, then it seems logical to understand that the participle construction is being used in exactly the same way.

A similar reasoning could equally be applied to the following example:

(17) Pa geseah he on swefne standan ane hlaedre fram eordan to heofenan, 7 Godes englas up stigende 7 nyper stigende on paere hlaedre

'Then he saw in a dream a ladder stand [ing] from earth to heaven, & God's angels going up & going down on the ladder'

(Heptateuch, 56: 12)

The referent of he, Iacob, presumably sees two events: that of the ladder standing between heaven and earth, and also that of the angels climbing and descending it. It has been pointed out elsewhere (for example, by Huddleston 1984: 384-387) that although coordinated elements generally tend to belong to the same syntactic category, this is not necessarily the case. (16) and (17) cannot be taken as proof, therefore, that VOSI and the NP + past participle string are syntactic equivalents. Nonetheless, they do suggest that if the first VOSI complement is clausal in nature (8), then presumably VOSende has been analysed in a similar fashion, with the NP viewed as the subject of a largely "verbal" participle.

There are indications that, in certain contexts at least, infinitive and participle constructions were becoming particularly close:

(18) Ond mon geseah hine blinde onlyhtende, ond hreofe claensian, ond laman gelacnian, ond deofol of mannum drifan, 7 deade aweccan, 7 windum stilnesse bebeodan, 7 dryum fotum gan ofer saes ypa, 7 opre wundro manega wyrcean

'And people saw him giving light to the blind, and cleans[ing] lepers, and driv[ing] the devil from men, & wak[ing] the dead, & order[ing] the wind to be still, & go[ing] over the waves of the sea with dry feet, & performing] many other wonders'

(Blickling, 124: 91)

The functional equivalence of VOSende and VOSI in this instance seems clear. The sentence constitutes a list of the miracles that Christ was seen to perform, within which giving light to the blind, expressed by the participle construction, seems to have exactly the same status as the other miracles, described by infinitive structures. From a syntactic point of view, the accusative pronoun hine (Christ), which is obviously the subject of each of the embedded infinitives claensian, gelacnian, drifan, aweccan, bebeodan, gan, and wyrcean, seems to stand in exactly the same relationship to the participle onlyhtende as to the infinitives. Apparently, two types of VP have been coordinated, with a shared subject, suggesting once again that the participle, much like the infinitive, functions at least as the head of a full VP of which hine is the subject.

Whatever its origins, therefore, there are signs, that VOSende could be used, certainly by later OE times, as a means of expressing the DPR of an event, and also that the participle was analysed, in some contexts at least, as a fully verbal element.

3. Participle inflection

There is a further parameter that needs to be taken into account, and which appears directly relevant to the question of the status of participles used in conjunction with perception verbs: that of the possibility for a participle to agree with the accusative NP.

3.1. Variation between inflected and uninflected participles

A closer examination of attested VOSende strings reveals that there are in fact two patterns, one in which the participle inflects, and one in which it does not, illustrated by (19) and (20) respectively:

(19) da geseah se Haelend Natanahel to him cumendne & cwaed be him, her is israhelisc wer on dam nis nan facn

'Then the Saviour saw Nathaniel coming towards him and said about him, here is an Israeli man, on whom there is no deceit'

(HC, West Saxon Gospels, 99)

(20) Opre daeg Iohannes geseah pone Haelend to him cumende & cwaed, her is Godes lamb, her is se pe ded aweg middaneardes synnae

'On another day John saw the Saviour coming towards him and said "here is God's lamb, here is the one who will put away the sin of this world"'

(HC, West Saxon Gospels, 57)

The participle in (19) carries the masculine accusative singular ending of the so-called "strong" adjective declension, -ne, to agree with the accusative NP, but not in (20). In Mitchell's terms, cumendne in (19) looks more "adjectival" in nature, whereas cumende in (20) appears more "verbal". In view of the other elements we have mentioned, one might be tempted to conclude that the NP + present participle string began life as a construction expressing the DPR of an entity, the accusative NP, modified by an adjectival participle. This construction would then have been grammaticalised, lexicalising the inference that, if one sees or hears an entity [x], doing [y], one in fact sees or hears the event [x does y]. Concomitently, the earlier construction in which the participle was felt to be a straightforward modifier of the accusative NP would have undergone a form of syntactic reanalysis, the latter becoming the head of a fully-fledged VP, of which the NP is henceforth the subject. This would fit the tendency mentioned by Boulonais (2011) for the use of participial constructions to be extended "from [the perception of] entities to [that of ] situations". Certainly, there are signs of variation, as one might expect if such a reanalysis were indeed taking place:

(21) peah pe he fram untrumum & unwisum preostum were gedered na geseah hine mon efre foroon eorne ne mid hatheortnesse onstyredne ne nenig man hine geseah swioe hlahendne ne nenig man hine geseah swioe grorniende

'Although he was hurt by weak and foolish priests, nevertheless no-one ever saw him stirred with anger, no man saw him laughing too much and no man saw him mourning too much'

(Blickling, 124: 91)

Whatever kind of perception is described by the first construction, with an inflected participle, appears to be identical to that evoked by the second, in which the participle remains uninflected. Examples such as (21) could be seen as a sign of hesitation on the part of individual speakers as to which form to use, as the change came to be diffused.

It should be possible to check the hypothesis by comparing the relative frequencies of inflected and uninflected participles over time. If it were correct, then we would expect to find proportionately fewer examples with inflected participles in later than in earlier OE. Unfortunately, however, the evidence proves less conclusive than might be hoped. The first problem concerns the limited number of examples available for comparison. In the relevant contexts, the present participle takes the so-called "strong" declension. Unfortunately for our purposes, the same form, -ende, is used for all genders, singular and plural, nominative and accusative, save the masculine accusative singular. Only a small proportion, therefore, of participle perception verb complements, those with a masculine singular noun, can usefully be compared. The total number of such occurrences (30) across all the OE texts in our corpus (9) is simply too small to be statistically significant.

Furthermore, distribution figures for those examples which are available do not suggest that a reanalysis of the type outlined above was taking place. It is true that in Aelfric's Homilies, a relatively late text (O3 by the Helsinki Corpus dating system (10)), the participle inflects in none of the 5 occurrences of VOSende with a masculine singular NP, whereas in the earlier (O2) Blickling text, the participle is inflected in 6 out of 7 similar cases. However, there is considerable variation here. Another O2 text, Bede, contains 3 examples with a masculine singular NP, all of which contain an uninflected participle. Overall, there is a slightly higher proportion of uninflected participles in the O2 texts (8 cases out of 14, or 57%) than in the O3 ones (8 cases out of 16, or 50%), but not the clear pattern one might expect to find if some kind of diachronic shift from more to less participle inflection were indeed taking place.

At the same time, the idea that the inflected participle is an adjectival modifier, whereas the uninflected form is more verbal in character, does not stand up to close scrutiny, as the following examples show:

(22) We gehyrap aefter disse aescan drihten andswariendne and pone weg his eardunge pus taecendne

'Let us hear the Lord answering and thus showing us the way to his house'

(HC, Aethelwold, 69)

(23) Forpon pe ic geseo mine Crist cigned ne me ond Paulus

'For I see my Christ summoning me and Paul'

(Blickling 130: 242)

Both taecendne in (22) and cigendne in (23), despite the accusative ending, are still used transitively, assigning accusative case (as well as the role of "theme") to what must be their internal arguments, pone weg and me ond Paulus respectively, in the manner of normal transitive verbs. As Denison (1993: 374) points out, this is a clear indication that the participles function not as adjectives but as the heads of full VPs.

But if the inflected participles cannot be analysed as adjectival forms, how are we to explain the alternation between inflected and uninflected uses? In order to have a better understanding of this phenomenon, we need to take a look at how present participles function in OE in contexts other than the complements of perception verbs.

3.2. Other uses of OE participles

There is one type of context in which the present participle is indeed used in a similar way to an adjective, illustrated by (24) and (25) below:

(24) Swa hwilc man swa me Apollonium lifigendne to gebringd, ic him gife fifti punda goldes

'Whichever man brings me Apollonius alive, I will give him fifty pounds in gold'

(Apollonius, 10: 16)

(25) ... paet ge eowerne eard mid waemnum bewerian wid onwinnendne here

'... that you defend your land with weapons against the attacking army'

(HC, Letter, 643-44)

The participles here are arguably more adjectival in nature in that each seems to modify an NP, and as such could easily permute with a straightforward adjective or some other form of modifying expression, unlike taecendne and cigendne in in (22) and (23). As far as I can tell, this kind of participle is always inflected. Otherwise, the overwhelming majority of participles occur in two types of context. They can form part of the beon/wesan periphrasis, as in (26):

(26) Aer oam fyrste waes se Halga Gast wunigende on dam apostolum

'Before then the Holy Ghost dwelt in the apostles'

(Aelfric, Homilies, 232: 23)

Alternatively, they can occur as what Denison (1993: 372) calls an "appositive" participle, which can simply be attached to any available subject in the context, as in (27):

(27) Iudei da, mid micelre stemne hrvmende. heoldon heora earan

'Then the Jews, crying with a loud voice, held their ears'

(Aelfric, Homilies, 46: 32)

This is Declerck's (1982: 15-18) "free adjunct". In both cases, the participle is presumably the head of a VP, and its subject almost exclusively a nominative NP. It therefore remains uninflected.

Free adjuncts, of course, should not be confused with the VOSende perception verb constructions that we have been looking at so far. The difference can be illustrated by examples like (28), where a perception verb is used with a free adjunct participle clause:

(28) [he] saegde paet he hine cneoht weasende gesawe

'He said that he saw it when he was a boy'

(Bede, 142: 6)

This example, from a purely syntactic point of view, is potentially ambiguous, in that the participle could have either he or hine as its subject. It is the context which informs us that weasende is understood to be attached to he. Hine refers here to a here, a 'sanctuary.' This syntactic ambiguity contrasts with the perception verb VOSende construction, where the participle can only be attached to the complement NP.

Of course, one of the consequences of adding the accusative -ne ending to the participle in VOSende perception constructions is to mark overtly the dependence of the participle on the object NP. It is possible that the accusative ending on the participle may simply have been felt, by some speakers at least, to be a means of marking the VOSende construction as a distinct syntactic entity. From this perspective, the different patterns observed in (19), (20), and (21) correspond not to a directional diachronic shift, from more frequent to less frequent, but rather to a case of synchronically competing grammatical options - a competition that appears to last for much of the OE period. In the relatively exceptional cases where the subject of the participle was a masculine accusative NP, speakers had two options: either to leave the participle with no distinctive inflection, as happened in the vast majority of other contexts, or to inflect it, feeling the construction to be different to the appositive one and marking it accordingly.

We could take this idea a little further, and envisage the possibility that we are looking at the emergence here, earlier than suggested in Lowrey (2010), of a largely grammaticalised pseudo-modifier construction. Simplifying somewhat11, this means that the NP and the present participle VP are indeed thought to form a constituent together, but an NP constituent, rather than a full clause. According to Declerck (1981: 155-157), "pseudo-modifier creation" involves the raising into object position of the subject NP of an embedded clause, turning the remainder of the clause, the participle VP, into a (syntactic) adnominal modifier. I shall not explore here the merits of Declerck's analysis of PDE structures, but simply retain the idea that agreement on the participle could be interpreted as an explicit marker of its pseudo-modifier function within the complex NP.

4. Latin influence

Other factors, too, need to be taken into account, the most important of which is probably the influence of Latin. To what extent does the grammar of a Latin original affect the syntactic choices made during the composition of an OE text? The question is particularly pertinent where VOSende is concerned, given the distribution of participle perception verb complements in OE.

In fact, VOSende with perception verbs is limited almost exclusively to prose texts in OE, those which tend to be translated from a Latin original. Callaway (1913: 228), pointing out that VOSende "is practically unknown in OE poetry," considers it not to be a native OE construction at all, but entirely the result of Latin influence. My own (admittedly limited) poetry corpus confirms Callaway's assessment of the distribution of VOSende. In Beowulf, the poetic texts from the Vercelli book, and the Seafarer, infinitive complements alone are used to descripe the direct perception of events. It appears very likely, therefore, that Latin influence did indeed play an important role here.

The exact extent of this influence, however, is not easy to determine. This can be demonstrated by directly comparing OE texts with their Latin originals. As an example, I shall take the late OE version of St. John's Gospel, translated more or less directly, as Bright (1904: xxvi) points out, from the Vulgate. The Latin original makes frequent use of a combination of an accusative NP and a present participle in the complement of direct perception verbs. The participle in Latin is systematically inflected, with the accusative -tem ending in the singular, and -tes in the plural, if the head noun of the NP is masculine. The syntactic forms found in the OE translations, however, tend to vary. Sometimes, the Latin participle construction is rendered by VOSI in OE:

(29) pa gesawon hig pone Haelend uppan paere sae gan, and paet he waes gehende pam scype; and hig him ondredon

'Then they saw the Saviour walk[ing] on the sea, and that he was near to the boat, and they feared him'

(St. John, 30: VI v19)

Vulgate: 'Cum remigassent ergo quasi stadia viginti quinque aut triginta, vident Jesum ambulantem supra mare, et proximum navi fieri, et timuerunt'

I shall assume, along with Callaway (1913) and Fischer (1989) that VOSI was a native OE construction.

On other occasions, the inflected Latin participle becomes a similarly inflected participle in OE, in what looks like a straightforward literal translation:

(30) he cwaed to him, paet eow beswicd? Gyf ge geseop mannes Sunu astigendnepaer he aer waes?

'And he said to him, what offends you? What if you saw the son of Man climbing back up where he was before'

(St. John, 35: VI vv61-2)

Vulgate: 'dixit eis: Hoc vos scandalizat? si ergo videritis Filium hominis ascendentem ubi erat prius?'

Sometimes, however, an inflected Latin participle is rendered by a participle in OE, but with no inflection:

(31) and he cwaed, pa he geseah pone Haelend gangende, Her is Godes lamb

'And he said, when he saw the Saviour com[ing], here is God's lamb'

(St. John, 5: I vv36-38)

Vulgate: 'Et respiciens Jesum ambulantem, dicit: Ecce agnus Dei'

It would appear, therefore, that although the Latin original almost certainly had a very strong influence on the OE version of the text, the translation is not necessarily a purely "slavish" one. While it is very likely that some, at least, of the variation between forms that has been described above can be attributed to Latin influence, those instances where the OE translation is not a direct "calque" from the Latin suggest that the translator retained at least a measure of native-speaker autonomy when choosing the syntactic form to use in the OE version.

5. VOSende in other contexts

It would be interesting to see if evidence for the existence of pseudo-modifier constructions in OE can be found in other contexts as well. Declerck (1981: 15961; 1982: 4-6) points out that pseudo-modifiers are not restricted to perception verb complements, but also occur in other contexts and in the complements of other verbs. In OE too, other verbs display complementation patterns similar to those of direct perception verbs. For instance, findan and gemettan, just like perception verbs, allow both infinitive and participle complements:

(32) Pa eode he eft ongean to oaes halgan martyres byrgenne, and funde his spere standan mid blode begleddod

'He then went again to the holy martyr's sepulchre, and found his spear stand[ing] stained with blood'

(Aelfric, Homilies, 452: 6)

(33) Paet cweartern we fundon faeste beclysed, and da weardas wioutan standende, ac we ne gemetton naenne wioinnan

'We found the prison shut fast, and the wardens standing without, but we found no one within.'

(Aelfric, Homilies, 572: 33)

(34) Hwaet se hlaford pa Garganus gegaderode micele menigu his in-cnihta, and done fearr gehwaer on dam westene sohte, and aet nextan hine gemette standan uppon oam cnolle paere healican dune

'Hereupon the master Garganus gathered a great many of his household servants, and sought the bull everywhere in the waste, and at last found him stand[ing] on the knoll of the high mountain'

(Aelfric, Homilies, 502: 12)

(35) Pa midpyssere bene beseah heo to daere stowe daer heo paet cild aer forlet, and gemette hit swa slapende swa heo hit aer gelede

'Then with this prayer she looked at the place where she had left the child earlier, and found it sleeping just as she had previously laid it'

(Aelfric, Homilies, 566: 16)

Semantically, of course, these structures give rise to much the same kind of ambiguity as VOSende with the perception verbs. Do we encounter an entity [x], doing [y] or rather the event of [x doing y]? Interestingly, just like in the complement of perception verbs, the participle could also be inflected to agree with a masculine singular object NP:

(36) pa gemette he hine paer swyde bitterlice wependne

'Then he found him there weeping most bitterly'

(HC, Gregory, 596)

(37) da com se wyrtweard on gewunelicre tide, 7 he pone peof paer on hege hangiendne funde

'Then the gardener came at the usual time, and he found the thief hanging there on high'

(HC, Gregory, 211)

If indeed the -ne inflection is an indication that some form of pseudo-modifier construction existed in OE with perception verbs, then examples such as (36) and (37) might also indicate that the pseudo-modifier had been grammaticalised to the extent that it was available for wider use, with a range of predicates.

6. Conclusions

It would appear from the above discussion that there are a number of reasons to believe that, in the Latin-influenced OE prose texts at least, what I have called the VOSende construction was used with perception verbs to mark the direct perception of events, in a manner similar to the VOSI construction. There is even evidence to suggest that the NP + present participle sequence had been reanalysed as a largely grammaticalised pseudo-modifier construction, as described by Declerck (1982). Some caution needs to be exercised, however, before any definitive conclusions can be drawn, given the relatively small sample which served as the basis for this study. Further research would be necessary to establish the exact status of VOSende in OE.

A number of important questions remain to be answered. One of these, of course, is that of the extent to which the OE participle constructions were influenced by the grammar of Latin. Some of the variation we have observed, at least, could simply be ascribed to hesitation on the translator's part as to how to render the Latin (inflected) participle construction: with a VOSI structure, with an inflected participle, as in Latin, or with an uninflected one. It remains to be determined whether, as Callaway (1913) suggests, VOSende was entirely a syntactic borrowing, or whether OE might have developed its own participle complements, at the same time perhaps as the beon/wesan + participle periphrasis spread. It would be interesting too to examine the relationship between the different kinds of structures used in OE as DPRs of events: VOSI, VOSende, but also Denison's (1993) V+I construction, in which the perception verb takes a subjectless infinitive complement. These are among the issues that will be examined in forthcoming papers.

doi: 10.1515/stap-2015-0002

REFERENCES

PRIMARY SOURCES

Aelfric, Homilies = Thorpe, Benjamin (ed.). 1844. Aelfric, The homilies of the Anglo-Saxon church, vol. I. London: Aelfric Society.

Aethelwold = The Benedictine rule, Helsinki Corpus.

Apollonius = Goolden, Peter (ed.). 1938. The Old English Apollonius of Tyre. Oxford: Oxford University Press.

Bede = Miller, Thomas (ed.). 1890. The Old English version of Bede's Ecclesiastical history of the English people. EETS OS 95, 96, 110, 111. London: Trubner.

Beowulf = Slade, Benjamin (ed.). 2003. Beowulf on Steorarume. Online edition, http://www.heorot.dk/

Blickling = Kelly, Richard (ed.). 2003. The BlicklingHomilies. London & New York: Continuum.

Dream of the Rood = Marsden, Richard (ed.). 2004. The Cambridge Old English reader. Cambridge: Cambridge University Press.

Gregory = Gregory's Dialogues, the Helsinki Corpus.

HC, Helsinki Corpus (Old English section) = The Helsinki corpus of English texts. 1991. Compiled by Matti Rissanen (Project leader), Merja Kyto (Project secretary); Leena Kahlas-Tarkka, Matti Kilpio (Old English); Saara Nevanlinna, Irma Taavitsainen (Middle English); Terttu Nevalainen, Helena Raumolin-Brunberg (Early Modern English). Department of English, University of Helsinki. http://www.helsinki.fi/varieng/CoRD/corpora/HelsinkiCorpus/

Heptateuch = Marsden, Richard (ed.). 2008. The Old English Heptateuch and Aelfric 's Libellus de Veteri Testamento et Novo (vol. I), EETS 330, Oxford: Oxford University Press.

Ine's Laws = Jebson, Tony (ed.). 1995. Laws of Aelfred and Ine. Online edition: http://www8.georgetown.edu/departments/medieval/labyrinth/library/oe/texts/prose/ laws.html.

Letter = Aelfric, Letter to Sigewaerd, the Helsinki Corpus.

Rushworth = Farman, Rushworth Gospels, the Helsinki Corpus.

Seafarer = Gordon, Ida L. (ed.). 1964. The Seafarer. London: Methuen.

St. John = Bright, James W. (ed.). 1904. The Gospel of Saint John in West Saxon. Boston & London: D. C. Heath.

St. Mary = Magennis, Hugh (ed.). 2002. The Old English life of St. Mary of Egypt. Exeter: University of Exeter Press.

Vercelli Book = Kemble, John M. (ed.). 1843-1856. The poetry of the Codex Vercellensis. London: Aelfric Society.

West Saxon Gospels = Anonymous, West Saxon Gospels, the Helsinki Corpus.

SECONDARY SOURCES

Boulonnais, Dominique. 2011. Non-finite complementation and perception verbs: A historical perspective. Paper delivered at CBDA-2 conference, Tours, 2011.

Callaway, Morgan. 1913. The infinitive in Anglo-Saxon. (Carnegie Institution Publications 169.) Washington: Carnegie Institution.

Declerck, Renaat. 1981. Pseudo-modifiers. Lingua 54. 135-163.

Declerck, Renaat. 1982. The triple origin of participial perception verb complements. Linguistic Analysis vol. 10(1). 1-26.

Felser, Claudia. 1999. Verbal complement clauses: A minimalist study of direct perception constructions. Amsterdam: John Benjamins.

Fischer, Olga. 1989. The origin and spread of the accusative and infinitive construction in English. Folia Linguistica Historica 8. 143-217.

Fischer, Olga & Ans van Kemenade & Willem Koopman & Wim van der Wurff. 2000. The syntax of Early English. (Cambridge Syntax Guides.) Cambridge: Cambridge University Press.

Huddleston, Rodney. 1984. Introduction to the grammar of English. Cambridge: Cambridge University Press.

Kirsner, Robert S. & Sandra A. Thompson. 1976. The role of pragmatic inference in sensory verb complements. Glossa 10. 200-240.

Lowrey, Brian. 2010. Quelques remarques sur la complementation des verbes de perception en moyen-anglais. In Jean-Charles Khalifa & Philip Miller (eds.), Perception et structures linguistiques: Huit etudes sur l'anglais. 91-112. Rennes: Presses Universitaires de Rennes.

Miller, Philip & Brian Lowrey. 2003. La complementation des verbes de perception en francais et en anglais. In Philip Miller & Anne Zribi-Hertz (eds.), Essais sur la grammaire comparee du francais et de l'anglais. 131-188. Vincennes: Presses Universitaires de Vincennes.

Mitchell, Bruce. 1985. Old English syntax. (2 vols.) Oxford: Clarendon Press.

Nickel, Gerhard. 1966. Die expanded Form im Altenglischen: Vorkommen, Funktion und Herkunft der Umschreibung 'beon/wesan' + Partizip prasens. Neumunster: Karl Wachholtz Verlag.

Quirk, Randolph & Charles L. Wrenn. 1957. An Old English grammar. (2nd edn.) London: Methuen.

Radford, Andrew. 1975. Pseudo-relatives and the unity of subject raising. Archivum Linguisticum 6. 32-64.

Smith, Jeremy. 1996. An historical study of English: Function, form and change, London & New York: Routledge.

Visser, Fredericus T. 1973. An historical syntax of the English language, Leiden: E.J. Brill.

Brian Lowrey

Universite de Picardie Jules Verne

(1) I am very grateful to Xavier Dekeyser for all his very helpful suggestions and comments on a first draft of this paper. Any remaining errors or inaccuracies are mine.

(2) For "Verb + Object and/or Subject + Infinitive," to take account of the fact that it can be difficult to determine whether the accusative NP is both the object of the higher verb and the subject of the infinitive, or merely the latter.

(3) Of course, it could be argued that perception and cognition can never be completely dissociated, and that all forms of so-called "direct" perception involve some degree of cognition. Nonetheless, there seems to be a clear distinction between perception, as expressed by a sentence such as He saw them begin the race and cognition as in He saw them to be hungry, where the meaning of see is close to that of understand, and may not even entail visual perception at all.

(4) As Xavier Dekeyser (p.c.) points out, the term "pseudo-modifier" is arguably something of a misnomer, in that there seems to be nothing "pseudo" about the construction. He suggests "covert modifier" instead, which might indeed be descriptively more accurate. However, given that Declerck's term has become widely accepted in the literature, I shall continue to use it here.

(5) The reader is referred to Declerck (1982) or to Felser (1999) for more extensive discussion of these properties in PDE.

(6) Of course, at a time before the grammaticalisation of the "progressive" form, VOSI constructions remain aspectually neutral. Position verbs such as standan or licgan appear frequently as infinitives in OE, in the complements of perception verbs, with an "imperfective" sense, and continue to do so until the end of the 18th century (see Lowrey 2010). I therefore add "[ing]" to the PDE translation of the infinitive here and in other examples to take into account the fact that OE infinitives in perception verb complements could be translated into modern English either by an infinitive or by an -ing form, according to the context. Sometimes the -ing form is necessary, as in (6), where the infinitive in modern English would tend to imply "go and stand", which is clearly not the intended meaning in the context.

(7) Interestingly, Fischer et al. (2000) seem to admit a "control" type analysis for VOSI with OE perception verbs, considering that subject NPs in VOSI structures "semantically ... are as much objects of geseah, as they are subjects of the infinitive" (2000: 221). This however, seems unlikely. As Kirsner and Thompson (1976) demonstrate, for PDE at least, we arrive at the conclusion that, in seeing the event [X do Y] we have necessarily seen the entity X, via a pragmatic inference, which can as such be cancelled. I have seen faith perform miracles does not imply, for example, I have seen faith. Since the same is apparently true of the corresponding VOSI constructions in other languages too, such as modern French or German, there seems to be no reason to assume that the same did not hold for OE. I shall return to this question in a forthcoming paper dealing with OE infinitive constructions.

(8) See Felser (1999: 90-123) for arguments in support of this view where PDE is concerned.

(9) The corpus used here is comprised of the texts listed as "Primary Sources" at the end of this paper.

(10) O2 corresponds to the period from 850-950, O3 to that from 950-1050 (presumed date of composition).

(11) More complete syntactic analyses of pseudo-modifiers in PDE are offered by Declerck (1981: 155; 1982: 10), and by Felser (1999: 64).