2. Old English

2.1. Introduction

Old English (hence OE) is a SOV language (the Germanic languages are OV/VO) with a head-final parameter: whatever the projection, the complement will precede the head (except for COMP whose complement is on its right).

(5)

It is also a V2 language: the finite verb raises to the second position of the sentence, whereas the topic on its left can be any element of the sentence. We shall first follow Pintzuk (1991)´s analysis on the OE sentence as an IP-V2 structure, i.e. the finite verb raises from V (its base-position) to I (or T) (the second position). That is what happens when we have a simple (or matrix) sentence, where TP is equivalent to IP.

(6)

(6.a)

Seuerus	se	casere	onfeng	micelne	dæl	Breotene...
Severus-NOM	the-NOM	emperor-NOM	received-IND.PRET	great-ACC	part-ACC	Brittany-GEN...

Severus the emperor received a great part of Brittany... (cobede,BedeHead:1.6.14.6)

(6.b)

In an embedded clause, we always have an IP-V2 structure. But for Pintzuk, there are two variants of this structure: an INFL-medial (7.a) and an INFL-final (7.b) ones.

(7)

(7.a)

(7.b)

(8)

...	þe	god	worhte	þurh	hine...
...	which	God	wrought	through	him...

... which God wrought through him... (Pintzuk 1991: 75)

INFL-medial structure:

(9)

[_C þe [_TP god_j [_T worhte [_VP t_j þurh hine [_V t_i]]]]]

INFL-final structure:

(10)

[_C þe [_TP god_j [_VP t_j þurh hine [_V t_i [_T weorhte [_PP þurh hine]]]]]]

And finally, for a small number of sentences (questions, sentences whose topic is a negative element or the time adverbs þa or þonne), we find a CP-V2 structure: the finite verb raises from V to T and to C, where Spec,CP is realized.

(11)

(12)

Ða	gewænde	seo	wydewe	ham...
Then	went-PRET	the-NOM	widow-NOM	back home...

Then, the widow went back home... (coaelive,ÆLS_[Eugenia]:144.277)

2.1.1. The different syntactic structures in contemporary English

Before dealing with the analysis of the Old English preterite present verbs, let us give the syntactic structures of some types of verbs, which we are to use in the course of our work. They all follow the same pattern: Subject + Finite Verb + Non finite Verb.

In contemporary English (CE), there are two types of verbs: lexical verbs and operator verbs. Among these different verbs, we can find control, raising or causative ones, or verbs whose complement is a TP.

A control structure is a structure whose infinitive clause has a PRO subject controlled by its antecedant:

(13)

(13.a)

They want to stay.

(13.b)

where the complement of the finite verb is a CP.

In Example (14.a), the complement of the finite verb is a TP.

(14)

(14.a)

They heard John cry

(14.b)

As for causative verbs, they are operator verbs: they are part of the event, but they are not the event; they introduce a notion of cause to the event. These are verbs such as make, have or cause in CE.

(15)

(15.a)

He made him cry.

(15.b)

At last, we can find raising structures. In CE, a raising verb lacks an external argument (i.e. a subject) and does not case-assign its complement, which generally is an infinitive clause. A raising verb is a lexical one such as seem or appear in CE. Moreover, modals can be found amongst raising verbs, and we assume modal verbs to be base-generated under T.

(16)

(16.a)

It seems that he understands her.

(16.b)

(17)

(17.a)

He seems to understand her.

(17.b)

(18)

(18.a)

He can understand her.

(18.b)

We kindly ask the reader to keep these structures in mind as we shall go back to them when analysing the preterite presents in parallel with causative verbs and raising structures.

2.2. The preterite present verbs

In Old English, there are two main classes of verbs: strong and weak verbs(note: ➳). The class of preterite present verbs is a particular one to which we shall add the anomal verb WILLAN.

Preterite presents are verbs which morphologically have a perfect form but a present meaning (for instance, ``I have got´´ in CE); WILLAN is an athematic verb, i.e. lacking a theme (note: ➳))

in *-mi: the first person singular for the indicative present was *-mi in Indo-European and in Germanic (bio-m en VA ``I am´´).

The verbs we are focusing on are the following (we shall follow Mossé´s classification from class 1 to class 6 (note: ➳) (Mossé (1945): 118-21):

• Class 1: WITAN know, AGAN (note: ➳) have, possess.

• Class 2: *DUGAN exist, be useful (note: ➳)

• Class 3: UNNAN grant, CUNNAN be able, can, know, ÐURFAN need, be necessary, DURRAN dare.

• Class 4: MUNAN remember, *SCULAN be obliged to, should.

• Class 6: *MOTAN be allowed to, can.

• Non classified: MAGAN be able to (note: ➳).

• Anomal verb: WILLAN want.

Present-day Modal verbs are to be found in this class of verbs. In the diachronic literature, it is generally assumed that modal verbs to have ``appeared´´ during the Middle English period. Before that, they were mere lexical verbs identical to the other verbs (see for instance Allen (1975) ou Roberts (1993)).

From a syntactic point of view, we are opposed to this statement given that the changes and evolutions within a language are progressive. So, the hypothesis underlying our analysis is the existence of a specific syntactic position for this class of verbs, as early as the Old English period. This position is different from the one assigned to strong and weak verbs which are base-generated under the head of VP.

Our analysis shall lead us to show that preterite present verbs are base-generated under a specific position we name v_Modal. Throughout our work, whe shall take into account the complex relations which exist between tense, mood and modality. To achieve this, we shall study their evolution through the use of different texts from different periods of time: Beowulf (cobeowulf in the examples to come) dating back to the 8^th century (≃ 680), Bede the Venerable´s Ecclesiastical History (cobede) dating from the 9^th-10^th centuries (890) and Apollonius of Tyre (coapollo) dating back to the 11^th century (1050).

2.3. Syntax and functional heads

We have just mentioned that OE is an SOV language with a V2 typology. Whether OE is a CP-V2 or an IP-V2 language, the functional heads C, T, (certainly) Neg and v, the head of transitivity, and the lexical head V exist. The analysis to come will allow us to add another head: another v, namely v_Modal still in the domain of vP.

2.3.1. The functional heads C, T and the lexical head V

We shall show the existence of such heads with examples taken from the OE corpus. The following example

(19)

and	cwæð:	þu	iunga	mann,	canst	ðu	þone	dom	mynra	dohtor	gifta?
and	said:	you-NOM	young-NOM	man-NOM,	know-IND.PRES	you-NOM	the-ACC	sentence-ACC	my-GEN	daughter-GEN	mariage gift-GEN?

and said: ``Thou young man knowest thou the condition of my daughter´s nuptials?´´ (coapollo,ApT:4.9.44; 1050)

illustrates Structure (11), and shows the existence of the head C. As for the following example,

(20)

Forðam	gif	hit	gewurðan	mæg,	ic	wille	me	bedihlian	on	eowrum	eðle.
Because	if	it-NOM	be	may-IND.PRES,	I-NOM	want-PRES	me	conceal	in	your-DAT	contry-DAT.

Therefore, if it may be, I will conceal myself in your country. (coapollo,ApT: 9.9.156)

it illustrates Structure (6.b) and shows the existence of the functional head T and the lexical head V.

Let us give some more examples to highlight what has just been said:

(21)

&	he	openlice	sæde	þæt	he	his	bebodum	hyrsumian	ne	wolde.
&	he-NOM	openly	said-PRET	that	he-NOM	his	orders-DAT	obey	NEG	want-PRET.

and he openly said that he did not want to obey his orders. (cobede,Bede_1:7.36.12. 292; 890)

(22)

þonne	sceal	he	hine	eaðmodlice	ahabban	from	onsægdnesse	þæs	halgan	gerynes...
then	shall-IND.PRES	he-NOM	him-ACC	respectfully	restrain	from	sacrifice	the-GEN	holy-GEN	sacrament-GEN...

then he must respectfully restrain himself from sacrifying the holy sacrament... (cobede, Bede_1:16.86.16.784)

(23)

Nolde	eorla	hleo	ænige	þinga	þone cwealmcuman	cwicne	forlætan.
NEG+wanted-PRET	warriors-GEN	protector-NOM	no-INSTR	thing-GEN	murderous guests-ACC	swift-ACC	go off.

Not for anything the protector of warriors let the murderous guests go off alive. (cobeowul,26.791. 678; 680)

(24)

Ne	meahte	ic	æt	hilde	mid	Hruntinge	wiht	gewyrcan.
NEG	might-PRET	I-NOM	in	fight-DAT	with	Hrunting-DAT	something-ACC	do.

With Hrunting I might not do anything in the fight. (cobeowul,51.1659.1375)

Examples (22) and (25.a) (=(19)) show the existence of the functional head C since we are dealing with CP-V2 structures,

(25)

(25.a)

and	cwæð:	þu	iunga	mann,	canst	ðu	þone	dom	mynra	dohtor	gifta?
and	said:	you-NOM	young-NOM	man-NOM,	know-IND.PRES	you-NOM	the-ACC	sentence-ACC	my-GEN	daughter-GEN	marriage gift-GEN?

and said: ``Thou young man knowest thou the condition of my daughter´s nuptials?´´ (coapollo,ApT:4.9.44; 1050)

(25.b)

Indeed, this example shows a CP-V2 structure, with movement of the finite verb from V to T, then from T to C. According to Pintzuk´s analysis (Pintzuk (1991), chapter 3), the way to know whether the structure is CP-V2 or IP-V2 is to note the position of the subject clitic: if it goes before the finite verb, it is an IP-V2 structure; but if it goes after the finite verb (as illustrated in our example), it is a CP-V2 structure (note: ➳). The constituant order on that kind of sentences (i.e. direct questions, sentences beginning with a negative constituant, or the time adverbs þa and þonne) is derived from the movement of the finite verb from T to C (Pintzuk (1991): 133).

Example (26.a) (=(20)) underlines the existence of the functional heads T and the lexical head V. This example is an IP-V2 structure.

(26)

(26.a)

Forðam	gif	hit	gewurðan	mæg,	ic	wille	me	bedihlian	on	eowrum	eðle.
Because	if	it-NOM	be	may-IND.PRES,	I-NOM	want-PRES	me	conceal	in	your-DAT	country-DAT.

therefore, if it may be, I will conceal myself in your country. (coapollo,ApT:

9.9.156)

(26.b)

For now, we will not go into details concerning this embedded structure. We shall go back to it further on in our work.

OE is generally assumed to be an asymmetrical language: V2 for matrix sentences, but not for embedded sentences since the position C is filled by a lexical complementizer. Pintzuk´s analysis shows that the structure of the OE sentence is IP-V2, both in matrix and embedded sentences (except for the small number of sentences we have just mentioned). She was able to prove the existence of INFL-medial structures for OE embedded sentences (with the finite verb raising from V to T, by analyzing the distribution of particles, pronouns, one-syllable adverbs, as well as the low frequency of the syntactic phenomenon called Verb (Projection) Raising (note: ➳). For embedded sentences, there is the coexistence of two structures: INFL-medial versus INFL-final structures. Within INFL-medial structures, the finite verb raises to T, which is in V2 position; in INFL-final structures, the finite verb also raises to T, but the order within TP is different for the head T is in final position.

Examples (23) and (24 show the existence of the functional head Neg, but not its syntax yet.

The existence of these functional heads allow us to introduce the theoretical framework we are working in: Chomsky´s Minimalist Program. Chomsky defines C and T as belonging to the Core Functional Categories of the sentence. C expresses force and mood and can be unselected (semantic selection); it selects T and has an EPP feature (i.e. a Spec). Concerning T, it must always be (semantically) selected, either by C (and then it possesses a bundle of ϕ features (features for person, gender and number), or by V,and then T is defective. Moreover T has an EPP feature and it selects V.

Then the structure of a sentence is (CP)-TP-VP, and we shall apply it to the OE examples to follow.

(27)

This structure can be applied to the following examples:

(28)

(28.a)

and	cwæð:	þu	iunga	mann,	canst	ðu	þone	dom	mynra	dohtor	gifta?
and	said:	you-NOM	young-NOM	man-NOM,	know-IND.PRES	you-NOM	the-ACC	sentence-ACC	my-GEN	daugther-GEN	marriage gift-GEN?

and said: ``Thou young man knowest thou the condition of my daughter´s nuptials?´´ (coapollo,ApT:4.9.44; 1050)

(28.b)

[_CP [_C canst [_TP [_Spec ðu [_T canst_i [_VP [_V canst_i þone dom mynra dohtor gifta ?]]]]]]]

(29)

(29.a)

Forðam	gif	hit	gewurðan	mæg,	ic	wille	me	bedihlian	on	eowrum	eðle.
Because	if	it-NOM	be	may-IND.PRES,	I-NOM	want-PRES	me	conceal	in	your-DAT	country-DAT.

therefore, if it may be, I will conceal myself in your country. (coapollo,ApT: 9.9.156)

(29.b)

[_CP Forðam [_C gif [_TP hit [_VP [_V gewurðan [_T mæg, [_TP [_Spec ic [_T wille [_VP [_V wille_i [_TP me bedihlian on eowrum eðle.]]]]]]]]]]]]

(30)

(30.a)

Ne	meahte	ic	æt	hilde	mid	Hruntinge	wiht	gewyrcan.
NEG	might-PRET	I-NOM	in	fight-DAT	with	Hrunting-DAT	something-ACC	do.

With Hrunting I might not do anything in the fight. (cobeowul,51.1659.1375)

(30.b)

[_CP [_C Ne meahte [_TP ic [_T meahte_i [_VP [_V meahte_i [_TP æt hilde mid Hruntinge wiht gewyrcan.]]]]]]]

At this stage, we shall not go any deeper into the base position of the preterite present verbs.

Examples (21), (23) and (24) give an illustration of the present and past forms of the preterite presents, but still we cannot say if these are indicative or subjunctive forms. Indeed, OE had two moods with different endings for each of them. The forms of Examples (19), (20) and (22) are in the indicative mood. If mood can be seen through endings on verbs, it would imply that there are two functional heads for both moods: one for the realis mood and one for the irrealis mood. We assume T to host [-irrealis] mood. But we shall go back to this point further on, once the syntactic position for preterite present verbs is defined.

2.4. Preterite presents and v_Modal

We have previously mentioned that preterite presents made up a class of verbs on its own, which was different from strong and weak. Why are they different? First, they are all defective (except WILLAN), i.e. they do not possess a complete system of conjugation (see Appendix A). Second, there is a vowel alternation between the infinitive form and the indicative present form which is not found with strong and weak verbs  (note: ➳).

Finally, preterite presents are the only verbs to have a specific ending for all the plural present forms: -on (witon, agon, cunnon, sculon, ..., whereas the lexical verbs have an -að (nimað ``they take´´, demað ``they decide´´, habbað ``they have´´) ending. They also are specified by the use of a dental suffix in -t(e), -d(e) and -ð(e) for the past singular forms, like weak verbs. Preterite presents are built both on strong and weak verbs (note: ➳).

These morphological differences are to be useful for our analysis. Indeed, if they have a specific morphology, the starting point of our analysis can be to grant them a specific syntactic position.

In Examples (19) to (24), we underlined the existence of the heads C, T and V, but we have purposely put aside Neg and the head for the preterite presents.

Let us now focus on them by questioning their base-position.

We propose there are two heads for the preterite presents verbs: V, for preterite presents used as lexical verbs, and a modal little v (i.e. v_Modal) for preterite presents followed by an infinitive complement.

Then, when a preterite present verb is followed by an object, it is base-generated under (lexical) V. It then merges to v and behaves the same way strong and weak verbs do: it is Structure (31). But when it is followed by an infinitive, it is to be considered as a semi-lexical verb (for grammaticalization is not completed yet). As a consequence, it is no longer base-generated under V but under v_Modal: it is Structure (32). Why assume such a position? All the instances of preterite presents verbs that have been found in the corpus are either to have a root reading or an epistemic reading (but we shall go back to this in further details in Section 2.8.2). We then have the following two verbal structures: (31) when the preterite present is lexical and (32) when it is semi-lexical:

(31)

(32)

In the examples to come, the preterite present is used alone (i.e. lexically):

(33)

No	hie	fæder	cunnon,	...
No	they-NOM	father-ACC	know-IND.PRES,	...

They know of no father, ... (cobeowul,42.1355.1119; 680)

(34)

Ac	sio	ðeod	þone	cræft	þæs	fiscaþes	ne	cuðe,	...
But	the-NOM	people-NOM	the-ACC	craft-ACC	the-GEN	fishing-GEN	NEG	knew-PRET,	...

But the people did not know the craft of fishing, ... (cobede,Bede_4:17.304.10.3076; 890)

(35)

ic	secge	ðe	to soðan	þone	forlidenan	man	ic	wille.
I-NOM	say-IND.PRES	he-NOM	in sooth	the-ACC	shipwrecked-ACC	man-ACC	I-NOM	want-PRES

I will say to thee in sooth that I desire the shipwrecked manJ. (coapollo,ApT:20.15. 427)

Let us take again Example (33) (now (36.a)) and give its structure where CUNNAN is first base-generated under V, then merges under v then T:

(36)

(36.a)

No	hie	fæder	cunnon,	...
No	they-NOM	father-ACC	know-IND.PRES,	...

They know of no father, ... (cobeowul,42.1355.1119; 680)

(36.b)

In most cases, the preterite present verb is followed by an infinitive (a strong or a weak verb, or another preterite present (in italics):

(37)

Nu	ic	eower	sceal	frumcyn	witan,	...
Now	I-NOM	you-ACC	shall-IND.PRES	ancestors-ACC	know,	...

Now I must learn your lineage, ... (cobeowul,10.251.203; 680)

(38)

Ic	to	sæ	wille	wið	wrað	werod	wearde	healdan.
I-NOM	to	sea-DAT	PRES	against	wrath-ACC	hosts-ACC	watcher-ACC	keep.

I shall go back to the sea to keep watch against hostile hosts. (cobeowul,12.318.257)

(39)

Wille

ic

asecgan

sunu

Healfdenes,

mærum

þeodne,

min

ærende,

aldre

þinum,

gif

he

us

geunnan

wile

þæt

we

hine

swa

godne

gretan

moton.

will-PRES

I-NOM

dire

son-DAT

Healfdene-GEN,

great-DAT

prince-DAT,

my-ACC

errand-ACC,

lord-DAT

your-DAT,

if

he-NOM

us-DAT

grant

wants-PRES

that

we-NOM

him-ACC

thus

kindness-ACC

address

shall-IND.PRES.

I will tell my errand to Healfdene´s son, the great prince your lord, if, good as he is, he will grant that we might address him. (cobeowul,13.344.285)

(40)

ac	wit	on	niht	sculon	secge	ofersittan,	...
but	the two of us-NOM	at	night-ACC	shall-IND.PRES	sword-ACC	forgo,	...

but we shall forgo the sword in the night, ... (cobeowul,22.681. 575)

(41)

...	þæt	he	wið	ælfylcum	eþelstolas	healdan	cuðe,	ða	wæs	Hygelac	dead.
...	that	he-NOM	against	foreigners-DAT	native throne-ACC	hold	could-PRET,	since	was-IND.PRET	Hygelac-NOM	dead-NOM.

... that he could hold his native throne against foreigners now that Hygelac was dead. (cobeowul,73.2367.1931)

(42)

he

scolde

eaðmodlice

for

heo

þingian,

þæt

heo

ne

þorfte

in

swa

frecne

siðfæt

&

in

swa

gewinfulne

&

in

swa

uncuðe

elþeodignesse

faran.

he-NOM

might-PRET

humbly

for

her-ACC

ask,

so that

she-NOM

NEG

needed-PRET

in

so

dangerous-ACC

journey-ACC

&

in

so

toilsome-ACC

&

in

so

incertain-ACC

pilgrimage-ACC

go.

that he might, by humble entreaty, obtain (of the Holy Gregory), that they should not be compelled to undertake so dangerous, toilsome, and uncertain a journey. (cobede,Bede_1:13.56. 6.521; 890)

(43)

Ic

gehyhte

&

wende,

þæt

wit

nu

hraðe

scoldon

ætgædre

in

ece

liif

gongan.

I-NOM

hoped-IND.PRET

&

believed-IND.PRET,

that

both-NOM

now

swiftly

should-IND.PRET

ensemble

dans

eternal-ACC

life-ACC

go.

I was in hopes that we should have entered together into life everlasting. (cobede,Bede_3:19.244.7.2499)

(44)

...þæt	he	swa	in	toweardnesse	ecelice	ricsian	mid	Criste	moste.
...that	he-NOM	so	in	hereafterr	eternally	reign	with	Christ-DAT	pouvait-PRET.

... so you will hereafter reign in Christ. (cobede,Bede_3:21.248.21. 2544)

(45)

...	ond	hy	ealle	þa	bliðe	mode	lustlice	healdon	woldon...
...	and	they-NOM	all-NOM	this-ACC	cheerful-INSTR	mind-INSTR	willingly	wanted-IND.PRET.

... and they would all most willingly observe (it) with a cheerful mind... (cobede, Bede_4:5.276.32.2812)

(46)

Gif	hit	nænge	þinga	to dæge	beon	mægge.
If	that-NOM	no-INSTR	thing-GEN	today-DAT	be	might-SUBJ.PRES.

If today that might happen with none of these things. (cobede,Bede_4: 12.290.20.2932)

(47)

...	þæt	nænig	ðara	onweardra	his	heortan	degolnessa	him	helan	dorste...
...	that	no-one-NOM	the-GEN	opponants-GEN	his	heart-GEN	secret-ACC	him-DAT	conceal	dared-PRET...

... none of the opponants dared to conceal the secret of his heart... (cobede,Bede_4:28.362.27.3643)

(48)

and	gif	ðu	þæt	ne	dest,	þu	scealt	oncnawan	þone	gesettan	dom.
and	if	you-NOM	this-ACC	NEG	do-IND.PRES,	you-NOM	shall-IND.PRES	understand	the-ACC	appointed-P.PASSE-ACC	sentence-ACC.

and if thou doest that not, thou shalt suffer the appointed doom. (coapollo,ApT:5.5. 71; 1050)

(49)

Hlaford

Apolloni,

gif

ðu

þissere

hungrigan

ceasterwaru

gehelpest,

na

þæt

an

þæt

we

willað

þinne

fleam

bediglian...

Lord-NOM

Apollonius,

if

you-NOM

these

hungry

citizens

help-IND.PRES,

no at all

this-DAT

only

that

we-NOM

will-IND.PRES

this-ACC

flight-ACC

conceal.

Lord Apollonius, if thou helpest these hungry citizens, we will not only conceal thy flight, ... (coapollo,ApT:9.18.164)

Let us illustrate Structure (32) with Example (48) (now (50.a)):

(50)

(50.a)

and	gif	ðu	þæt	ne	dest,	þu	scealt	oncnawan	þone	gesettan	dom.
and	if	you-NOM	this-ACC	NEG	do-IND.PRES,	you-NOM	shall-IND.PRES	understand	the-ACC	appointed-P.PASSE-ACC	sentence-ACC.

and if thou doest that not, thou shalt suffer the appointed doom. (coapollo,ApT:5.5. 71; 1050)

(50.b)

When we are dealing with negation, the adverbial negative particle ne (following the terminology used in the diachronic literature) immediately precedes the finite verb,

(51)

...	no	ðu	ymb	mines	ne	þearft	lices	feorme	leng	sorgian.
...	no	you-NOM	about	me-GEN	NEG	needs-IND.PRES	body-GEN	comfort through food-DAT	longer	trouble.

... no longer will you need trouble yourself to take care of my body. (cobeowul,16.448.374; 680)

(52)

ða	hine	Wedera	cyn	for	herebrogan	habban	ne	mihte.
after	the-ACC	Weather-Geats-GEN	people-NOM	because of	fear of war-DAT	keep	NEG	might-PRET.

After that the country of the Weather-Geats might not keep him. (cobeowul,16.459.386)

(53)

Ða	ne	meahte	he	eaðelice	þa	unstillnesse	onfallendra	mengu	aberan...
Then	NEG	could-PRET	he-NOM	easily	the-ACC	agitation-ACC	oppressing-GEROND-GEN	crowd-GEN	bear...

he could no longer bear the crowds that resorted to him... (cobede,Bede_3:14.216.32.2220; 890)

(54)

ac	he	ne	mihte	hine	þar	findan	on	ðam	flocce.
but	he-NOM	NEG	could-PRET	him-ACC	there	find	in	the	troops-DAT.

but he could not find him in the company. (coapollo,ApT:13.9.232)

Moreover, this adverbial particle merges with that specific class of verbs (to be more accurate, it merges with AGAN, WILLAN and WITAN (note: ➳)):

(55)

and	hi	noldon	me	þa	agifan.
and	they-NOM	NEG+wanted-IND.PRET	me	her-ACC	restore.

and they would not restore her to me. (coapollo,ApT:50.10.534; 1050)

(56)

Nat	he	þara	goda	þæt	he	me	ongean	slea.
NEG+knows-PRES	he-NOM	these-GEN	tools-GEN	that	he-NOM	me-DAT	against	strike-SUBJ.PRES.

He knows no good tolls with which he might strike against me. (cobeowul,22.681.574; 680)

(57)

...	þæt	he	þa	weorþuncge	Eastrena	on riht	ne	heold	ne	nyste.
...	what	he-NOM	the-ACC	observance-ACC	Easter-GEN	correct-ACC	NEG	respected-IND.PRET	nor	NEG+knew-PRET.

... what he imperfectly understood in relation to the observance of Easter. (cobede, Bede_3:14.206.1.2087; 890)

What is striking is that merging only happens with this specific class of verbs (even if it is not compulsory), but not with strong and weak verbs (except with wesan and habban).

Let us sum up their differences:

1. They have a specific relationship with negation since ne can merge with the preterite present.

2. They are morphologically different from the other classes of verbs (as far as their endings are concerned). The plural form of the indicative present is not the strong or weak verbs; furthermore, there is a vowel alternation between the infinitive and the present indicative forms.

3. The preterite present verbs can generally take

   • an object NP (either noun or pronoun), like any other class of verbs,

   • a VP where there is an infinitive verb (for strong and weak verbs, it belongs either to TP or CP).

   as a complement.

When an infinitive follows a lexical verb (strong or weak), it belongs to TP or CP, according to the type of verb one is dealing with. We shall analyze this in more detail in the next section (note: ➳)

(58)

(58.a)

(58.b)

(59)

Þa	het	he	hraðe	his	þegnas	hine	secan	&	acsian.
Then	ordered-IND.PRET	he-NOM	immediately	his	soldiers-ACC	him-ACC	seek	&	call.

Whereupon he sent some soldiers to make a strict search after him. (cobede,Bede_1:7.34.25.280)

(60)

...	þæt	seolfe	he	ne	blinneþ	mærsian	&	weorþian	a	butan ende...
...	that	self-NOM	he-NOM	NEG	ceases-IND.PRES	glorify	&	celebrate	always	ever.

... or rather never ceases to celebrate... (cobede, Bede_5:20.474.6.4756)

Let us take Example (59) (now (61.a)) and give its structure (which will be identical for Example (60)).

(61)

(61.a)

Þa	het	he	hraðe	his	þegnas	hine	secan	&	acsian.
Then	ordered-IND.PRET	he-NOM	immediately	his	soldiers-ACC	him-ACC	seek	&	call.

Whereupon he sent some soldiers to make a strict search after him. (cobede,Bede_1:7.34.25.280)

(61.b)

All the given examples allow us to say that preterite presents are base-generated under V when they are lexical. But when they are semi-lexical, they are base-generated under v_Modal. We further assume that they are raising verbs with a structure very close to the CE causative structures. They do behave like operators: they are part of the event of the sentence but they are not the event (because of that, they are said to be ``causative´´). Moreover, they do not have an external argument (i.e. a subject, the visible subject is the subject of the infinitive verb) and they do not case-assign their complement. Subject and object are assigned θ-roles by the infinitive. Still, agreement is visible between the subject and the preterite present.

2.4.1. The functional head v

At the beginning of this chapter, we have defined three functional heads (C, T and Neg (even if we have not defined the syntactic position of the latter)) by means of examples. We are now going to define a fourth functional head – little v – which also belongs to the CFCs of a phase. In Chomsky (1998): 15, v is a ``light´´ verb, head of transitive structures. It must always be selected by the functional category T. In its turn, it selects a VP and an NP (i.e. the external argument of v). It has an EPP feature as C and T do. Moreover, within the minimalist framework, C and v define the strong phases of a derivation. Let us point out that fact in OE.

(62)

...	þa	cliopode	heo	hi	hire	to	mid	liðere	spræce...
...	then	called-PRET	she-NOM	her-ACC	her-DAT	to	with	gentle-DAT	speech-DAT.

then she called her to her with gentle speech... (coapollo,ApT:2.14.24)

The structure of this sentence is the following:

(63)

In this structure, there are two strong phases vP and C. The speaker first chooses elements which correspond to a (lexical or grammatical) category from his/her lexicon. Both the verb and its complements are generated under VP which has semantically been selected by v. This is why V merges with v the head of transitivity of the (strong) phase. Then, the subject in Spec,vP satisfies the EPP feature of TP. Next, an additional Spec is adjoined to vP for the internal argument.

The speaker then chooses the following elements in his/her lexicon:

• V = cliopode [+past,-irréel,+sg3],

• NP_Subject = heo [+sg3,+fem,+nom],

• NP_Object = hi [+sg3,+masc,+acc],

• PP = hire to [+sg3,+fem,+dat],

• PP = mid liðere spræce [+dat].

Next, the chosen components combine with one another:

(64)

At this point of the derivation, it can still crash because the tense features of the verb and T are not satisfied yet. As a consequence, the derivation cannot be spelled out. So, v is to merge with T to satisfy the uninterpretable tense features of T, and the subject in Spec,vP is to raise to Spec,TP to satisfy the ϕ features of T. Once satisfied, the tense features and the ϕ features of the subject are erased/cancelled. But still, there is no Spell-Out since TP is not a strong phase. As we have just said, T is semantically selected by C. Merge of the time adverb þa under Spec,CP triggers movement from v to C. The derivation is now over and can be spelled out.

So, we have just underllined the existence of v, another functional head. But we shall question Neg later in our work.

Within Halle et Marantz´s Distributed Morphology (DM), and more particularly in Marantz (1999), v has the following features:

1. It creates a verb.

2. It provides event semantics.

3. It provides agentive semantics for agentive constructions.

4. It merges with an external argument.

5. There is an Agree relation with the object.

According to Marantz, Features 1 to 3 go together, i.e. the semantic content of v acts on the semantic content of the sentence. For a specific v, we can have Features 1 to 3 without Features 4 and 5, as we can have Features 4 and 5 without Features 2 and 3.

Following our hypothesis, v_Modal has Features 1, 2 and 3: it creates the preterite present which is a semi-lexical verb (that we can consider as an ``operator´´; it provides event semantics (it adds up information to the relation set by the non finite verb); and it might provide agentive semantics (yet, can a raising verb provide agentive semantics?). But v_Modal does not have Features 4 and 5 because the external argument merges with the non finite verb, and because there is no Agree relation with the infinitive verb to which it does not assign case.

2.5. Existence of two positions for the preterite presents

Up to now, we have postulated the existence of two syntactic positions for the preterite presents:

1. The position V when the preterite present is direct transitive (followed by an NP), indirect transitive (followed by a PP) or bitransitive (followed by two objects).

2. the position we have named v_Modal when the preterite present is transitive (the complement is an infinitive, i.e. a VP). This latter position is lower than T.

2.5.1. Infinitive structures

For the sake of our analysis, it is necessary to define what an infinitive structure is and what the differences are when the finite verb is either lexical or semi-lexical.

The analysis we carried out in the past section allows us to say two syntactic positions exist in OE for preterite present verbs depending on their lexical or semi-lexical use. When lexical, the syntactic position is V; when semi-lexical, the syntactic position is v_Modal. The following example illustrates the first position. But in both cases, the subject is generated under Spec,vP (but when vP is tacit, the subject is under Spec,VP).

(65)

(65.a)

Ac	sio	ðeod	þone	cræft	þæs	fiscaþes	ne	cuðe,	...
But	the-NOM	people-NOM	the-ACC	craft-ACC	the-GEN	fishing-GEN	NEG	knew-PRET,	...

But the people did not know the craft of fishing, ... (cobede,Bede_4:17.304.

10.3076; 890)

(65.b)

When a preterite present does not have a NP complement, it is VP where the infinitive verb is.

Let us have a close look at the infinitive structures where there is either a lexical or a semi-lexical verb, so that we can define the differences between these two types of verbs.

In CE, infinitive clauses have the following structures, with a lexical verb with TO ((66.b)) or without TO ((67.b)); or with a modal verb (68.b)):

(66)

(66.a)

They want to stay.

(66.b)

(67)

(67.a)

They heard John cry.

(67.b)

(68)

(68.a)

Mary can swim.

(68.b)

The infinitive verb is the complement for the modal and it is base-generated under VP. As for the other types of infinitives where the finite verb is lexical, they are clauses, that is a complex sentence with a finite and non finite verb. Concerning the lexical V, it does not raise to T, but it is the tense affix that lowers to V. This is Affix Hopping.

Now, what about Old English? Do infinitive clauses belong to a VP (which is our hypothesis)? Before analyzing these clauses, let us first focus on infinitives following a finite lexical verb.

From the OE period, we can find two types of infinitive clauses: the ones introduced by TO (Examples (69) to (75)), and the ones which are not introduced by TO (Examples (59) and (60)).

(69)

ræd	eahtedon	hwæt	swiðferhðum	selest	wære	wið	færgryrum	to	gefremmanne.
plan-ACC	soughtt-IND.PRET	what-NOM	strong-hearted-DAT	the best-NOM	was-SUBJ.PRET	avec	awfulness-DAT	TO	follow-DAT.

sought a plan, what would be best for strong-hearted men to do against the awful attacks. (cobeowul,8.171.138)

(70)

No	þæt	yðe	byð	to	befleonne...
No	that-NOM	easy-NOM	is-IND.PRES	TO	flee-DAT.

That is not easy to flee from... (cobeowul,32.1002.835)

(71)

and	þæt	gefremed	man	gewilnode	to	bediglianne.
and	that-ACC	stranger-ACC	man-ACC	wanted-PRET	TO	conceal-DAT.

and he wanted to conceal that stranger. (coapollo,ApT:1.14.13)

(72)

willaþ	þæt	gewrecan	gif	we	magon,	þeah	we	beotiaþ	to	∅.
want-IND.PRES	that-ACC	revenge	if	we	can-IND.PRES,	even if	we-NOM	threaten-IND.PRES	TO	∅.

and desire, if we can, to take revenge, and (if we are unable), we nevertheless threaten to do so. (coblick, HomS_10_[BlHom_3]:33.127.447)

(73)

&	het	him	to	gelangian	þa	ylcan	Gotan.
&	ordored-IND.PRET	him-DAT	TO	send	the-ACC	same-ACC	Goths-ACC.

& he ordered him to send for the same Goths. (cogregdH,GD_1_[H]:10.80.10.794)

(74)

Eft

þæs

on

mergen

het

se

manfulla

dema

þa

eadigan

Agnen

him

to

gefeccan.

Again

this-GEN

on

morning-ACC

ordered-IND.PRET

le-NOM

wicked-NOM

judge-NOM

the-ACC

saint

Agnes-ACC

him-DAT

TO

seek.

The morning after, the wicked judge ordered Saint AGnes to be sought for again. (coaelive,ÆLS[Agnes]:91.1774)

(75)

Ða

cwæð

to

him

oþer

of

hys

leorningcnihtum,

Drihten,

alyfe

me

ærest

to

farenne

&

∅

bebyrigean

minne

fæder.

Then

said-IND.PRET

to

him-DAT

other-NOM

of

his

disciples-DAT,

Lord-NOM,

allow-SUBJ.PRES

me

before

TO

go-DAT

&

∅

bury

my-ACC

father-ACC.

Another of his disciples then said to him: ``may the Lord allow me first to go and bury my father´´. (cowsgosp,Mt_[WSCp]:8.21.462)

If we look closely at these examples, we can notice that when an infinitive clause is introduced by TO, the infinitive can be marked dative, which does not seem to be the case when TO is absent. We noticed this in Example (72): there is an ellipsis of the non finite verb, which would show that TO becomes a functional head (we shall explore that later in our work). On the contrary, in Example (73), the infinitive is unmarked. We can then wonder what this TO syntactically stands for: is it the well-known functional head in CE, is it a preposition governing the dative case, or is it a semi-lexical item which will lead us to say that TO follows the same syntactic path preterite presents do? In view of the preceding examples, we can consider TO as a semi-lexical preposition (i.e. undergoing grammaticalization) governing the dative case (see Roberts, Roussou (2003)). Yet, the syntactic representation of this element would be the functional head T having a [+dat] feature which is checked out by the [+dat] feature of the infinitive (if this feature is marked).

Let us now illustrate what we have just said with the following three examples: (76) where TO is not present, (77) where TO precedes the non finite verb, and (79.a) where PRO the subject of the infinitive is controlled by the subject of the sentence.

(76)

Þa	het	he	hraðe	his	þegnas	hine	secan	&	acsian.
Then	ordered-IND.PRET	he-NOM	immediately	his	soldiers-ACC	him-ACC	seek	&	call.

Whereupon he sent some soldiers to make a strict search after him. (cobede,Bede_1:7.34.25.280)

(77)

and	þæt	gefremed	man	gewilnode	to	bediglianne.
and	this-ACC	stranger-ACC	man-ACC	wished-PRET	TO	conceal.

and the perpetrated crime sought to conceal. (coapollo,ApT:1.14.13)

Whose structure is,

(78)

As for the example

(79)

(79.a)

No	þæt	yðe	byð	to	befleonne...
No-NOM	easy-NOM	is-IND.PRES	TO	flee-DAT.

That is not easy to flee from... (cobeowul,32.1002.835)

(79.b)

In Examples (80) to (84),we give again instances of preterite presents followed by an infinitive, in Examples (80) to (82), they are followed by an infinitive clause where the non finite is realized; but in Examples (83) to (84), there is an ellipsis of the non finite verb.

(80)

ne	mihte	snotor	hæleð	wean	onwendan.
NEG	could-PRET	wise-NOM	hero-NOM	sadness-ACC	change.

nor might the wise warrior set aside his woe. (cobeowul,8.189.153)

(81)

ic	geðristlæhte	þæt	ic	dorste	þis	weorc	ongynnan.
I-NOM	assumed-PRET	that	I-NOM	dared-PRET	this-ACC	work-ACC	begin.

I assumed I dared to begin this work. (cobede,BedePref:4.10.163)

(82)

and	hi	noldon	me	þa	agifan.
and	they-NOM	NEG+wanted-IND.PRET	me	her-DAT	restore.

and they would not restore her to me. (coapollo,ApT:50.10.534)

(83)

cwæð	heo:	Wilt	ðu,	wit unc	abidde	ondrincan.	Cwæð	ic:	Ic	wille	∅...
said-IND.PRET	she-NOM:	want-IND.PRES	you-NOM,	me and you-NOM	call-SUBJ.PRES	drink.	said-IND.PRET	I-NOM:	I-NOM	want-PRES	∅...

she said, ´Would you like me to call for something to drink?´ - ´Yes,´ said I, ´and am very glad if you can.´ (cobede,Bede_5:3.392.32.3921)

(84)

unc	sceal	worn	fela	maþma	gemænra	∅,	siþðan	morgen	bið.
we-DAT	shall-IND.PRES	a lot-NOM	many	treasures-GEN	precious-GEN	∅,	since	morning-NOM	is-IND.PRES.

many of our treasures will be shared when morning comes. (cobeowul,55.1782.1472)

2.5.2. Causatives structures

In contemporary English, causative verbs are operator verbs: they are neither lexical nor auxiliary verbs, they are part of the event but they are not the event. A causative verb is a v having the features 1 to 5 we have already mentioned in Section2.4.1: it creates a verb, provides event and agentive semantics, merges with an external argument and has an Agree relation with an object. Let us give some examples of causative verbs again in CE, then let us focus on OE to see if we can find such instances.

(85)

He had them eat the cake.

(86)

He made him cry.

The structure of (86) is,

(87)

In CE causative structures, the object of the finite verb is a VP. Is it the same in OE? We have considered the instances of the verb don ``cause, do´´:

(88)

(...)	doð	foroft	drymen	&	wiccan	on	heora	scincræfte,	to	beswicenne...
(...)	cause-IND.PRES	very often	magicians-NOM	&	soothsayers-NOM	by	their	craft-DAT,	TO	deceive-DAT...

Very often with their craft, the magicians and soothsayers are deceiving... (coaelhom,ÆHom_18:91.2543)

(89)

Þone

oðerne

dæl

he

dyde

∅

gehealden

mid

him

to

bebyrgenne

æfter

his

forðsiðe.

This-ACC

other-ACC

speech-ACC

he-NOM

caused-PRET

∅

follow

with

him-DAT

TO

bury-DAT

after

his-DAT

death-DAT.

he had this other speech be followed before him for his burial once he died. (coaelive,ÆLS_[Basil]:123.531)

(90)

and	dydon	on	wæter	wanhalum	to	þicgenne,	...
and	caused-IND.PRET	opn	water-ACC	stagnant-DAT	TO	drink-DAT,	...

and they made (him) drink this stagnant water, ... (coaelive,ÆLS_[Oswald]:200.5496)

(91)

Ond	heo	leornunge	godcundra	gewreota	hire	underþeodde	dyde	to	bigongenne,	...
And	she-NOM	knowledge-DAT	holy-GEN	scriptures-GEN	her-DAT	subjected-P.PASSE-DAT	caused-PRET	TO	attend-DAT,	...

And she obliged those who were under her direction to attend so much to reading of the Holy Scriptures, ... (cobede,Bede_4:24.334.16.3354)

the last examples have the following structure which applies to the other examples (even (89) where TO is not visible):

(92)

Considering these examples (they are not numerous in the corpus we are using, the infinitive structure of this causative verb is TP in Old english, and not VP like in CE. They are then different from preterite present verbs which have a VP complement.

Let us go back to preterite presents. The (finite) preterite present is followed by an infinitive in (80) to (82). In Examples (83) to (84), the finite verb is alone but there is the ellipsis of the infinitive verb. We will now analyse the structure of these verbs and show that the infinitive complement is indeed a VP. We will then be able to differenciate causative verbs from preterite present verbs and claim that causative verbs are VP whereas preterite presents are v_Modal (according to the analysis in Marantz (1999), our v_Modal only has three features out of five, two of them dealing only with semantics). Before starting our analysis, let us give some examples we have already mentioned:

(93)

willaþ	þæt	gewrecan	gif	we	magon,	þeah	we	beotiaþ	to	∅.
want-IND.PRES	that-ACC	revenge	if	we	can-IND.PRES,	even if	we-NOM	threaten-IND.PRES	TO	∅.

and desire, if we can, to take revenge, and (if we are unable), we nevertheless threaten to do so. (coblick,

HomS_10_[BlHom_3]:33.127.448)

(94)

...	þæt	heo	his	wedenheortnisse	gestilden,	ac	heo	ne	meahton	∅.
...	that	he-NOM	his	madness	eased-SUBJ.PRET,	but	he-NOM	NEG	could-IND.PRET	∅.

... did all he could to assuage the madness of the unfortunate man, but could not prevail. (cobede,Bede_3:9.184.32. 1853)

(95)

unc	sceal	worn	fela	maþma	gemænra	∅,	siþðan	morgen	bið.
we-DAT	shall-IND.PRES	a lot-NOM	many	treasures-GEN	precious-GEN	∅,	since	morning-NOM	is-IND.PRES.

many of our treasures will be shared when morning comes. (cobeowul,55.1782.1472)

The first step of our analysis is drawing a parallel between Example (93) and Examples (94) to (95). So far, the examples where the infinitive clause was introduced by TO have showed us that TO tended to be a functional head, that is a semi-grammatical preposition. Example (93) confirms this statement: with the ellipsis of the infinitive, TO is considered as a functional head (we assume it is T); if TO had been a preposition, the ellipsis of the infinitive would not have occurred. We then suggest to draw a parallel between this structure, where we assume TO to be base-generated under T, and Structures (94) to (95). This analysis will allow us to show that v_Modal is the syntactic position where preterite presents are base-generated when they are followed by an infinitive. So, when there is an ellipsis of the non finite verb, both heads hosting TO and a preterite present are ``functional´´. Their syntactic structures are slighty identical: the structure corresponding to TO + V is,

(96)

and the structure corresponding to Pret.Pres. + V (note: ➳) is,

(97)

According to this structure, preterite present verbs seem to be raising verbs: the semantic subject of the non finite verb raises to Spec,TP to become the grammatical subject of the sentence, where Agree is visible between the subject and preterite present.

Let us go back to Examples (93) and (94) (now (98.a) and (99.a)) and illustrate them.

(98)

(98.a)

willaþ	þæt	gewrecan	gif	we	magon,	þeah	we	beotiaþ	to	∅.
want-IND.PRES	that-ACC	revenge	if	we	can-IND.PRES,	even if	we-NOM	threaten-IND.PRES	TO	∅.

and desire, if we can, to take revenge, and (if we are unable), we nevertheless threaten to do so. (coblick,

HomS_10_[BlHom_3]:127.446)

(98.b)

(99)

(99.a)

þæt	heo	his	wedenheortnisse	gestilden,	ac	heo	ne	meahton	∅.
...	that	he-NOM	his	madness	eased-SUBJ.PRET,	but	he-NOM	NEG	could-IND.PRET	∅.

... did all he could to assuage the madness of the unfortunate man, but could not prevail. (cobede,Bede_3:9. 184.32.1854)

(99.b)

Let us refine our analysis of preterite presents to stress their syntactic difference. We are going to look at:

1. examples displaying the structure MD + INF + OBJ-ACC,

   (100)

   Mæssepreost

   sceal

   habban

   mæsseboc

   and

   pistelboc,

   and

   sangboc

   and

   rædingboc

   and

   saltere

   and