previous
TOC
next
The Syntactic Evolution of Modal Verbs in the History of English

3. Middle English

3.1. A few generalizations

Let us begin by a quotation,

``Middle English““ (in French) is the language which was spoken and written in Great Britain between 1150 and 1550 (...). These dates are only convenient mile stones: normally, there is no continuity solution for the evolution of a language. But if one compares the set of features caracterizing texts written during the ``Middle English““ period to those from the preceding period, one is struck by somes constant differences: a) vowels tend to get a standard tone (when they stil exist) e in (any) endings; b) a deep change and a great simplification of the inflection were determined by this levelling; c) as a consequence, there is a tendency (that has already begun in OE) to have recourse to analytic constructions or to prepositions instead of cases; d) finally, the lexicon has borrowed words from the French (consequences of the Norman conquest) or the Scandinavian languages (further to the Danelaw colonization) (note: ).

Compared to OE, the structure of the ME sentence is somewhat different, and tends to become the structure we know in CE. The main changes are:

Whereas in Old English, the order object-verb was very frequent, in particular in subordinate clauses and when the object was a pronoun, in Middle English this order became gradually less common, and ceased to show a correlation with clause type. (...) Another, probably related, Middle English change affected the position of particles relative to the verb. (...) such elements were also often preverbal in Old English. In the course of the Middle English period they gradually came to be restricted to postverbal position. In this case too, however, the older order continued to be used every now and then until the end of the Middle English period. (...) A further Middle English change involving verb position is the decline of the so-called `Verb-Second“ rule. (...) Verb-Second rapidly declined in the course of the last part of the fourteenth and in the fifteenth century, and saw a revival in the literary language in the sixteenth century (Fischer, van Kemenade, Koopman, van der Wurff (2000): 82-3).

We can also quote Kroch & Taylor:

Pintzuk (1991, 1993 and 1995) have shown that the transition from INFL- final to INFL-medial word order was a long-term trend characterizing the entire Old English period, so that its disappearance in Early Middle English can be taken as a continuation of Old English development rather than a break with it (...). [Early Middle English texts] exhibit all three of the base orders that have been proposed for Old English: INFL-final with an OV verb phrase, INFL-medial with an OV verb phrase, and the modern order – INFL-medial with a VO verb phrase (Kroch, Taylor (2000):132-3).

These three base orders are the consequence of the different influences the ME dialects have been under. Still according to Kroch and Taylor,

Although we are not primarily concerned with the historical sociolinguistic dynamics that established the ME dialects, the sociolinguistic history of population contact and diffusion which underlie them is a matter of considerable interest (...). Specifically, we will see that the northern dialect of English most likely became a CP-V2 language under the extensive contact it had with medieval Scandinavian, contact that resulted from the Danish and Norwegian population influx into the North of England during the late OE period (...). The linguistic effect of this combination of population movement and population mixture was extensive (Kroch, Taylor (1997) : 298-9).

Moreover,

The difference in the position to which the verb moves in different languages leads to subtle but clearly observable differences in the shape and distribution of verb second clauses. Most strikingly, while all V2 languages exhibit verb second order in main clauses, the two sub-types [IP-V2 and CP-V2] differ in the availability of this word order in subordinate clauses. The CP-V2 languages allow verb second order only in those embedded clauses that in some way have the structure of matrix clauses, either because the complementizer position is empty or because there is an additional complementizer position below the one that introduces the subordinate clause (the so-called `CP-recursion“). (...) The IP-V2 languages, on the other hand, show V2 order in a broad range of subordinate clauses. (...) We will further see that the southern dialect of ME preserves the V2 syntax of OE, despite having become, unlike OE, overwhelmingly I-medial and VO in basic order. In striking contrast to the southern dialect, however, the northern dialect of ME appears to have developed the verb-movement syntax of a standard CP-V2 language and hence to be similar in its syntax to the modern Mainland Scandinavian (Kroch, Taylor (1997) : 297-8).

To sum up, we could say that the ME syntax still displays the OE syntactic order (V2 language), as well as the CE order, and that V2 rapidly declined from the end of the 14th century.

3.2. Preterite presents

The members of this specific class of verbs are the same compared to OE, apart from one member. Like in Chapter 1, they all follow Mossé“s classification (Mossé (1945))  (note: )

MON was not part of the class in OE, and is no part of it in Early Modern English nor in CE.

We shall mainly base our analysis on the following texts: the Ormulum (?c1200; East Midlands: south-west of Lincolnshire), Ancrene Riwle (1225-1230; West Midlands) and Chaucer“s The Canterbury Tales (1380-91; East Midlands: Londres) (note: ), as well as other texts from the ME corpus (Kroch, Taylor (2000)) if we require more specific examples.

In the last chapter, we showed the existence of the functional heads C, T, Neg and v. Our analysis led us to add some more: a second Neg head, vModal for deontic preterite presents and Mood for epistemic preterite presents (which also hosts irrealis).

Similarly to OE, we also find the following structures: the ones where there is a lexical preterite present (i.e. followed by a direct object) and the ones where there is a semi-lexical preterite present (i.e. followed by an infinitive) base-generated under vModal).

(266)
& tohh he it nowwhar funde žęr, Ne wollde he it nęfre cunnenn, ...
& although he-SUJET it-OBJET nowhere found-PRET there, NEG voulut il-SUJET cela-OBJET jamais savoir, ...

and though he did not find it anywhere, he did not want to know, ... (CMORM, I,26.318)

(267)
ant alle aʒen hire in an eauer to halden.
and all-SUJET should her-OBJET in an ever TO observe.

and everyone should observe it always in the same way. (CMANCRIW,I.44.34)

(268)
... ne nenoteš naut his wit as mon ach to donne...
... NEG knows-NEG+PRES NOT his mind-OBJET as homme-SUJET ought TO do...

... he does not know his mind as a man should... (CMANCRIW,II.48.447)

(269)
ʒef ha aʒen tobeon feor from alle worldliche men hwet hu ancren aʒen to hatien ham & schunen, ...
if they-SUJET ought TO+be far from all wordly men-PL. what how cloistered-SUJET PL. ought TO call them-OBJET & teach, ...

if they have to get away from all the worldly men, this is how the cloistered men ought to call and teach them, ... (CMANCRIW,II.67.721)

(270)
... ase hwa se žus seide, Ich nolde forto žolien deaš ...
... as as so thus said-PRET, I-SUJET NEG+would FOR+TO suffer death-OBJET ...

as he then said, I did not want to suffer death... (CMANCRIW,II.76.889)

(271)
deaš me ach to fleon ase forš se me mei-PRES wiš uten sunne.
death-OBJET one-SUJET ought TO as in as one-SUJET may with out sun.

you should flee from death as one may flee before the sun. (CMANCRIW,II.85.1029)

(272)
& tęrfore hafe icc turrnedd itt Inntill Ennglisshe spęche, Forr žatt I wollde bliželiʒ Žatt all Ennglisshe lede Wižž ęre shollde lisstenn itt...
& therefore have-PRES I-SUJET translated-P.PASSE it-OBJET until English language, for that I-SUJET would-PRET perfectly that all English people-SUJET with respect should listen it-OBJET...

and I have therefore translated it into English because I wanted all the English people to listen to it respectfully... (CMORM,DED.L113.33)

(273)
Forži ach že gode habben eauere witnesse...
Therefore ought the good-SUJET have ever witness-OBJET...

Therefore good should always have a witness... (CMANCRIW,II.56.543)

Like in OE, we find occurrences of preterite presents used lexicaly, i.e. followed by an object either direct or indirect. Yet, if we have a closer look all the different texts, we can notice that only a small number of these verbs are used this way.

3.3. State and syntactic changes

Let us recall the OE sentence when a preterite present is displayed. We shall only mention the main functional heads. SOV Sentences can be CP-V2 (direct questions, sentences introduced by a negative element or by the adverbs ža and žonne) as in Example (274) or IP(TP)-V2 as in Example (275)

(274)
Tree 79
(275)
Tree 80

Whenever no preterite present is displayed, the structures remain identical.

Before going any further in our analysis, let us go back to the structures of preterite presents in OE when they are lexical: Pret.Pres. + THAT-clause and Pret.Pres. + (FOR)TO-infinitive. In both cases, the complement of the preterite present is a CP.

When the complement is a CP-infinitive, it has a PRO subject controlling the subject of the sentence. The preterite presents are then control verbs which assign a θ-role to the subject (unlike raising verbs which do not have an external argument and do not assign external θ-role: they are one argument-verb) and to its object, the internal argument of V. The subject is case-marked nominative in all personal constructions, and mainly case-marked dative in impersonal ones.

(276)
Tree 81

Lexical preterite presents are base-generated under V, they assign a θ-role to the subject base-generated under Spec,vP. The first phase is over. The subject then raises to Spec,TP to satisfy the EPP features of T and the preterite present satisfy its tense feature under T. C then assigns nominative case under government to the subject, and Agree is done through T: the second phase is over and can be spelled out.

We have also underlined that these very verbs were semi-lexical when followed by an infinitive and they were also raising ones. Moreover, the grammatical subject of the preterite present is the semantic subject of V. They are not base-generated under V but under vModal, the functional head we introduced in the last chapter. We recall their syntactic structure:

(277)
Tree 82

In late ME, we underlined that we could already talk about grammaticalisation (note: ) since we found examples taken from Denison (1993) displaying epistemic-reading preterite presents. Moreover, we stressed the syntactic position Mood (above T within the domain of CP) corresponding to these preterite presents. Indeed, C is a periphrastic functional head which gives some more information about the sentence, about the relation established between a subject and its predicate. As far as epistemicity is concerned, it allows the speaker to voice an opinion reflecting his/her understanding or knowledge about this pre-established relation. So, Mood expresses the additional information brought by the speaker. Nevertheless, the epistemic preterite present is generated under T, and the phonological level, the [-deont] feature of Mood will ``hop““ to T, following the CE Affix Hopping (note: ). On a DM point of view, the morphemes of T and Mood merge, the two nodes remaining distinct.

As for the deontic-reading preterite presents, they are generated lower than T, under vModal within the domain of vP. We then obtain the following structure in OE, and specially in ME:

(278)

CP - MoodP - TP - vModalP - (vP) - VP - (CP) - (TP) - (VP).

In ME, things change syntactically. The examples to come illustrate the major value change of the parameters: ME is now an SVO language. NP subjects and objects do not have morphological endings for person anymore; the features impoverish, triggering a change of value of the parameters, hence the structure becomes rigid. Yet, some structures remain V2. Then this syntactic competition (SOV-SVO sentences), the appearance of the dummy DO, the grammaticalisation and the epistemic readings of the preterite presents (note: ) are partially responsible for the setting of grammaticalisation. As a consequence, we shall name modal a preterite present which underwent grammaticalisation, i.e. a lexical or semi-lexical item becomes a grammatical, or functional, item.

The following examples show the ME SVO structure. As for questions or sentences beginning by a negative element (or focus), their structure is CP-V2 as we underlined in the previous chapter.

(279)
Žeʒʒ shulenn lętenn hęželiʒ Off unnkerr swinnc lef brožerr.
They-SUJET shall let honourably off their labour beloved brother.

Beloved brother, they shall let their labour go. (CMORM,DED.L53.21)

(280)
Ža seʒʒde Zacariass žuss Till Godess enngell sone; Žurrh whatt maʒʒ icc nu witenn žiss Žatt itt me muʒhe wurrženn?
Then said Zacharias thus till God“s angel soon; through what may I-SUJET now know this-OBJET That it-SUJET me-OBJET might become?

Then Zacharias said to God“s angel: ``Alors Zacharie dit ą l“ange de Dieu : ““ Through this, I can then know it, what might become of me?"` (CMORM,I,4.156)

(281)
ʒiff žu willt wurrženn borrʒhenn, Acc nohht onn ane wise žohh, Swa summ že boc-SUJET uss-OBJET kižežž.
If you-SUJET will be born-P.PASSE, But NOT on one way then, So some the book us becomes known-PRES.

If you want to be backed in a different way, as the book makes it known to us. (CMORM,I,172.1423)

(282)
Heo teacheš al hu me schal beoren him wišuten, ...
She-SUJET teaches-PRES all how one-SUJET INDEF shall bear him-OBJET out, ...

She teaches all about how to bring him forth, ... (CMANCRIW,I.42.17)

(283)
... žet ich mote blisfulliche grete že in heouene.
... that I-SUJET might blissfully greet him-OBJET in heaven.

... that I might greet him in heaven. (CMANCRIW,I.72. 277)

(284)
ʒef ani seiš wel ošer deš wel, ne maʒen ha nan loken židerwart wiš richt echʒe Of god heorte.
if any-SUJET says-PRES well or does-PRES well, NEG may he-SUJET none look thither with right eye of good heart.

if anyone says or does something good, he cannot look there with a rightful eye. (CMANCRIW,II.157.2133)

(285)
ant žu schalt finden in ham gretunges fiue.
and you-SUJET shall find in him greetings-OBJET PL five.

and you shall find five greetings in him. (CMANCRIW,I.74.290)

(286)
Že vres of že Hali Gast, ʒef ʒe ham wulleš seggen, seggeš bifore Vre Lauedi tiden.
The hours-SUJET PL of the holy Spirit, if you-SUJET them-OBJET will say, say-IMP before our Lady time.

If you want to say the hours of the Holy Spirit, say them before our Lady“s time. (CMANCRIW,I.74.297)

(287)
And with this swerd shal I sleen envie.
And with this sword shall I-SUJET kill envy-OBJET.

And I shall kill envy with this sword. (CMASTRO,662.C2.24)

(288)
For certes, al the sorwe that a man-SUJET myghte make fro the bigynnyng of the world...
For certain, all the sorrow that a man-SUJET might make from the beginning of the world...

For sure, all the sorrow that a man might cause from the beginning of the world... (CMCTPARS,291.C2.133)

(289)
For certes, the derke light-SUJET that shal come out of the fyr that evere shal brenne shal turne hym al to peyne that is in helle...
For certain, the dark light-SUJET that shal come out of tha fire that ever shall burn shall transform him-OBJET all to pain that is-PRES in hell...

For sure, the dark light that shall come out of the everburning fire shall transform him into all the pain that exists in hell... (CMCTPARS,291.C2.137)

(290)
And in this same wise maist thow knowe by night the altitude of the mone or of brighte sterres.
And in this same way may you-SUJET know by night the altitude-OBJET of the moon or of bright stars-PL.

By night, you can likewise know the altitude of the moon or the bright stars. (ID CMASTRO,669.C2.211)

(291)
... for thou scholdest knowe that the mowynge of schrewes, whiche mowynge the semeth to ben unworthy, nis no mowynge;
... for you-SUJET should know that the ability-SUJET of tyrants-PL, which ability you-OBJET seems-PRES TO be unworthy-OBJET, NEG+is-PRES no ability-OBJET;

... for you should know that the capacity of tyrants, which seems to be unworthy to you, is not one; (CMBOETH,448.C1.398)

If the examples have an epistemic reading, their syntactic structure is,

(292)
Tree 83

and if they have a deont reading, the structure is,

(293)
Tree 84

In a Section to come, we shall go back to these different structures to analyze them more accurately.

3.4. Grammaticalisation

The word grammaticalisation can be defined as followed (Roberts, Roussou (2003): 6):

The term grammaticalisation was first introduced by Meillet (1912) to describe the development of new grammatical (functional) material out of `autonomous“ words. [...] As Hopper and Traugott (1993: 1-2) point out, the term `grammaticalisation“ can be used to either describe the framework that considers ``how new grammatical forms and constructions arise““ or ``the processes whereby items become more grammatical through time““.

Another definition would be (from Matthews (1997)):

The process by which, in the history of a language, a unit with a lexical meaning changes into one with grammatical meaning(note: ).

Concerning modal verbs, the preterite presents, which were lexical and semi-lexical items, have become grammatical items, i.e functional words. This process of grammaticalisation can be seen through the syntax of the ME sentence, as a consequence of a morphological impoverishment (especially the loss of the irrealis endings) and the setting of new values to the existing parameters (such as the grammaticalisation of the preposition TO or the gradual appearance of the dummy DO) during that period. Nevertheless, this process took place gradually, as can be shown with the examples we are to analyze.

3.4.1. Consequences on the preterite presents

Until the ME period, preterite presents are semi-lexical verbs belonging to a specific class of verbs, hence behaving differently in syntactic terms. Whether epistemic or deontic, they are now raising verbs. Epistemic modals are base-generated under T and governed by MoodP; root modals are still base-generated under vModal and then raise to T, since they are grammatical, unlike lexical verbs.

In ME, they can still be used as lexical verbs (specially WILLEN, DURREN and CUNNEN), but the great majority of them tend to be followed by an infinitive which either has an epistemic or deontic reading.

This statement is very interesting: if there is an ambiguity between epistemic and root readings, it implies that these verbs indeed grammaticalize. (note: )

We assume the ambiguous readings reflect the grammatical change of these verbs: from lexical items to grammatical items. The impoverishment of the verbal morphology, the rarer use of V2, a more rigid syntax and its relation to negation (which now does not precede the finite verb), and semantics (there are some shifting meanings for the preterite presents from the OE to the ME periods) are the reflection of the change of status of these verbs.

This grammaticalisation can also be seen morphologically: even if some forms still bear number and person, and specially 2nd person singular present or past, these verbs now only have a present and a past form. The present form is different from the form for lexical verbs since they possess the 3rd person singular in the present, as for the past form, there are no more differences between the irrealis endings.

How can grammaticalisation be visible? Within the phase theory, the chosen items from the lexicon are already inflected and they satisfy a set of features with functional heads. Within Halle, Marantz (1993)“s Distributed Morphology, each morpheme corresponds to one functional head whose uninterpretable features are identical to the morphemes“s. So, when there is impoverishment, some features of the morphemes disappear, reducing then the competition between lexical items. Let us take an example: in OE, we have canst, the present form (2nd person singular) of the preterite present CUNNAN. If we use a tree structure to represent it, we obtain:

(294)
Tree 85

But we can also provide another morphological analysis of it. We can consider each morpheme on its own and say that a functional projection correponds to each of them. The structure we then get is not very different from the previous one. The main change is the status of the root <cαn> which is now one of the morpheme constituting the paradigm canst. The structure would be:

(295)
Tree 86

The V-1 notation is borrowed from Jensen (1990); it underlines that can is now a root, and it is the sum of all the morphemes which builds up the verb canst. This approach is different from Minimalism where the verb is chosen from the lexicon fully inflected. Yet this approach is not incompatible with it for it can reflect the grammaticalisation of the preterite presents in ME. Indeed, if these verbs are to be considered as paradigmas made of a root and a set of features, this phenomenom (i.e. grammaticalisation) leads to the loss of the endings and the vowel alternation (in the present indicative) of the preterite presents, that is, the morphosyntactic features of these lexical items are deleted. And it further leads to the deletion of some other features from other morphemes (what is called delinking in DM: when this delinking occurs, it entails the delinking of features which are dependent on them (the change is always from the more marked to the less marked value).

Let us take again the example žu canst and compare it with its contemporary equivalent you can (this form will be found in ME). We can note that the morphological mark of second person has disappeared and there is no longer vowel alternation (in OE, CUNNAN → can; in ME CUNNEN → can but CAN → can in CE).

(296)
Tree 87

We shall refer to this structure further on in our work, but correct it because after grammaticalisation modals are no longer semi-lexical but they are grammatical items.

Let us now go back to Examples (279) to (291), which allow us to underline the following important points for our analysis.

  1. We have an SVO surface order, even if morphology is still visible on modals for the 2nd person singular and plural (Examples (285), (286) and (290)). Our hypothesis is that these morphological marks on modals (tense, person and number) and thou/ ʒe subjects (2nd person singular and plural) do not have any influence on the SVO structure.

  2. We still find CP-V2 structures illustrating OE SOV structures. They can either be introduced by a negative element (Example (284)) or a topic (Examples (287), (290) and (291)).

  3. In Examples (280) and (286) (but there are many others), we can question the strusture: is it SOV, hence OE, or SVO, for the object is found inbetween the subject and the finite verb, as we have considered so far that the object is the internal argumant of vP.

  4. Finally, Examples (279), (282), (283), (288) and (289) display the ME SVO structures where cases are no longer morphologically visible (specially NOM and ACC, which freezes the structure), as person and number are on modal verbs.

With the coexistence of these two structures in ME, we also question the existence of the fuctional head reflecting this grammaticalisation. This point has been introduced in the previous chapter where we stessed the functional head vModal for semi-lexical preterite presents. This functional head still exists in ME, as Mood which has been introduced for epistemic modals and irrealis in OE. These two heads reflect two main types of modality in English: epistemic modality and deontic modality, even if according to Denison (1993), epistemic readings of modals are less common than deontic readings. We could sum up this established fact saying that in the tree structure everything under T, that is vModal, is semi-grammaticalised and everything above T, that is Mood, is grammaticalised. T then marks the boundary between these two states, which means that the modal is or is not within the semantic scope of T. If it is not, the evaluation of the predicative relation is at the tense of the utterance; then it has an epistemic reading (belonging to the irrealis sphere) and the morphological past has no time value. But if tense scopes on the modal, it is deontic.

In the next sections, we shall analyze the preterite presents in relation to modality, underlining that both epistemic and deontic modals have the same syntactic structures, the only difference being that epistemic modals are base-generated under T and deontic modals under vModal. For deontic modals in particular, we shall draw a parallel with causative structures. We shall also sketch the analysis of the grammaticalisation of the preposition TO (followed by an infinitive) to relate it to the grammaticalisation of modal verbs. And we shall also go back to the notion of perfective aspect, as well as to the syntactic structure of causative verbs.

3.5. Infinitive structures: general points

In the previous chapter, we saw that according to whether we were dealing with preterite present or lexical verbs, the infinitive structure was different. Is it the same in ME? Let us recall that OE infinitives have a particular ending -an, whatever the verb. In ME, we still find this ending -en, but a great number of infinitives are now bare for the morphological mark has disappeared. But this morphological loss(note: ) is also responsible for the grammaticalisation of preterite presents, among other things.

Concerning lexical verbs, we can note different types of structures:

  1. The finite verb is followed by an inflected infinitive with TO,

    (297)
    that alle schrewes ne ben worthy to han torment?
    that all tyrants-SUJET NEG are-PRES worthy TO have torment-OBJET?

    that all tyrants are unworthy to be tormented? (CMBOETH,448. C2.421)

  2. The finite verb is followed by an infinitive introduced by FORTO,

    (298)
    ... he shall cumenn efft To demenn alle žede-OBJET, & forr to ʒeldenn iwhillc mann Affterr hiss aʒhenn dede.
    ... he-SUJET shall come often TO judge all-OBJET country-OBJET, & FOR TO give back each-OBJET man-OBJET after his own death.

    ... he shall often come to judge all the country and give each man back after his death. (CMORM,DED.L171.39)

    (299)
    Prudence, his wyf, (...) bisoghte hym of his wepyng for to stynte.
    Prudence, his wife, (...) implored-PRET him-OBJET of his tears FOR TO stop.

    His wife Prudence implored him crying to stop. (CMCTMELI,217.C1b.10)

  3. the finite verb is followed by an infinitive without TO,

    (300)
    Whan Prudence hadde herd hir housbonde avanten hym of his richesse and of his moneye, ...
    When Prudence-SUJET had-PRET heard-P.PASSE her-SUJET husband-SUJET boast him-OBJET of his wealth and of his money, ...

    When Prudence had heard her husband boast about his wealth and money, ... (CMCTMELI,232.C2.601)

    (301)
    ... Ne munnde he nęfre letenn himm Žurrh rodepine cwellenn;
    ... NEG remembered he-SUJET never let him-SUJET through the torment of the cross kill;

    ... He never remembered that he had let him be executed on the cross; (CMORM,I,68.612)

  4. Finally, infinitive structures whose finite verb preceding them is let,

    (302)
    žach he seo žt heo him mis paiʒe he let hire ʒete iwurden.
    although he-SUJET sees-PRES that she-SUJET him-OBJET dislikes he-SUJET lets-PRES she-SUJET yet be.

    although he realizes he dislikes her, he lets her be as she is. (CMANCRIW,II.161.2223)

Let us now give examples of infinitive structures introduced by modals:

(303)
But nathelees, men shal hope that every tyme that man falleth, be it never so ofte, that he may arise thurgh Penitence...
But nevertheless, men-SUJET PL shall hope that every time that one-SUJET falls-PRES, is-PRES it-SUJET never so often, that he-SUJET may arise through penitence...

Nevertheless, men shall hope that whenever one falls (be it rarely), he can arise through penitence... (CMCTPARS,288.C2b.21)

(304)
First a man shal remembre hym of his synnes;
First a man-SUJET shall remember himself of his sins-PL;

A man must remember his sins; (CMCTPARS,290.C1.75)

(305)
... whil that iren is hoot, men sholden smyte...
... while that iron-SUJET is-PRES hot, one-SUJET should smite...

... one should strike while the iron is hot... (CMCTMELI,219.C1.82)

(306)
And therfore, er that any werre bigynne, men moste have greet conseil and greet deliberacion.
And therefore, before that any war-SUJET begins-PRES, one-SUJET must have great advise and great deliberation.

And therefore, before any war begins, a meeting should be held to deliberate. (CMCTMELI,219.C2.90)

(307)
Forr whase mot to lęwedd follc Larspell off Goddspell tellenn, He mot wel ekenn maniʒ word Amang Goddspelless wordess.
For whoever must to simple people sermon-OBJET of Gospel tell, he-SUJET must well develop many word-OBJET among Gospel words-PL.

For, whoever must say a sermon from the gospel to the ignorant people has to explain many words from it. (CMORM,DED.L53.16)

(308)
Ža Goddspelless alle žatt icc Her o žiss boc maʒʒ findenn, Hemm alle wile icc nemmnenn her Bi žeʒʒre firrste wordess.
The gospels-PL all that I-SUJET here in this book may find, them-OBJET all FUTUR I-SUJET name here by their first words-PL.

All the gospels I can find here in this book, I will name them using their first words. (CMORM,DED.L335.65)

(309)
For uh an schal halde žuttere...
For each one-SUJET shall hold other...

Because each one must hold the other... (CMANCRIW,I.44.26)

(310)
& heo schal habbe leaue to gladien hire fere, ...
& she shall have love TO rejoice her power, ...

& she shall have (enough) love to rejoice her power, ... (CMANCRIW,II.57.552)

(311)
and ye shal fynde refresshynge for youre soules.
and you-SUJET shall find refreshment-OBJET for your souls-PL.

and you shall find refreshment for your souls. (CMCTPARS,288.C1b.10)

The common change to all these verbs is that the infinitive can take en or be bare.

With AGEN, the infinitive structure is identical to that of lexical verbs for it is also introduced by TO, which reflects the grammaticalisation of this verb (note: ), and the parallel that can be drawn with the grammaticalisation of TO.

These different examples (modals and lexical verbs) are interesting from a morphological point of view since we have a ``competition““ between the two structures:

  1. on the one hand, modal + INFINITIVE-en and verb + TO+ INFINITIVE-en, and

  2. on the other hand, modal + INFINITIVE-∅ and verb + TO + INFINITIF-∅.

In Section 2.5.1, we analyzed the difference between the infinitive structure of modals and lexical verbs:

- [vModal Preterite present [vP subject Infinitive]] and,

- [VP VLexical[CP [C ∅ [TP PRO [T (TO) [VP Infinitive]]]]]] ou [VP VCausative [TP (TO) Infinitive]].

So, do these structure change when the infinitives become bare, i.e. uninflected? In other words, does this morphological loss have an influence on syntax, and more accurately on the preterite presents and TO?

In Roberts, Roussou (2003): 112-9, the authors claim that the loss of the subjunctive and the inflection of infinitives plays an important role in the grammaticalisation of TO (the modal content of the subjunctive ending being transferred to TO), by supposing that mood features are now realized in a higher position, and not by the ending anymore. So, they consider the verbal forms of infinitive constructions introduced by TO as ``subjunctives““: they are morphologically identical and, according to the authors, the increase of this kind of structure in ME is due to the decrease of subjunctive clauses displaying a marked morphology. Even if we do not follow their analysis of the TO particle, the morphological loss of the -en ending of infinitives and the -e/-en endings of the subjunctive is part of the grammaticalisation of some items such as the preterite presents or TO. As seen previously, the preterite presents and TO particle behave syntactically the same in relation to the ellipsis of the non finite verb. We can find this fact in ME.

(312)
and gief he hadde werred wiš god alse že deuel him to eggede...
and if he-SUJET had-PRET fought-P.PASSE with god also the devil-SUJET him-OBJET TO approached-P.PASSE ...

and if he had fought God, and that he had been appraoched by the devil ... (CMTRINIT,195.2699)

(313)
to speke that thou woldist not ∅...
TO speak that you-SUJET would NOT ∅...

say what you would not ... (CMAELR4,3.64)

3.5.1. Causative structures

Compared to what we have found in OE, causative verbs in ME do not have the same structure: they now have a VP complement like preterite presents and not a TP complement anymore. It would imply that these verbs do have a different syntactic position. In OE, we showed these verbs were lexical Vs; in ME, we assume they are semi-lexical verbs for their structure is identical to the CE one, that is operator verbs, and that they are vs. We voice the hypothesis they are now base-generated under v. The grammaticalisation of causative verbs would justify the grammaticalisation of deontic modals:

(314)
(314.a)

causative verb: lexical (V) ⟶ semi-lexical (v),

(314.b)

deontic modal: semi-lexical (vModal) ⟶ grammatical (vModal: the position grammaticalised; in DM, vModal and T fuses.

Let us give some examples.

(315)
žu habbe heo idon mid že licome...
you-SUJET have-PRES them-OBJET do with the flesh...

you have them do with flesh... (CMLAMBX1,21.242)

(316)
uor uirtue makež wynne heuene, and onworži že wordle...
or virtue-SUJET makes-PRES win heaven, and be worth the world...

or virtue makes us grant paradise and deserve the world... (CMAYENBI,84.

1630)

(317)
(they sholden) at the last maken hem lesen hire lordshipes...
(they should) at the last make them-SUJET lose their powers-PL...

at last, they should make them lose their powers... (CMCTMELI,231.C1.532)

Syntactically, it implies that they do not have the same position: they change from lexical to semi-lexical verb, whereas modals change from semi-lexical to grammatical verb, as shown in Example (314) (for deontic modals).

Thanks to these different examples (except the ones illustrating the causative verbs ``have““ and ``make““), we have shown that modals and TO syntactically behave the same way with respect to the ellipsis of the non finite verb. Then, there exists a parallel between the evolution of the preterite presents and the evolution of TO. This is why Roberts and Roussou“s idea is attractive, and we can apply it to the preterite present verbs: with the loss of the infinitive and subjunctive endings, these verbs can undergo a grammatical change. They grammaticalise, and this can be seen through a great number of forms of should and would which belong to the irrealis sphere, i.e. conditional (note: ). But we shall deal with this later on in our work.

3.6. Modality

3.6.1. Epistemicity and deonticity

Contrary to OE, epistemic and deontic readings are easier to notice in ME because of the morphological impoverishment and semantic shifts of the preterite present verbs which trigger their grammaticalisation. The notions charaterizing them (possibility, necessity and obligation) are clearer. Moreover, some verbs like SCULEN and WILLEN are more and more used to express the future (which was already the case in OE but to a lesser extent), and the conditional.

Let us take some examples of deontic modals,

(318)
ʒho ža shollde ben žurrh Godd Off Haliʒ Gast wižž childe.
you-SUJET then CONDITIONAL be through God of Holy Spirit with child.

you should bear a child from the Holy Spirit. (CMORM,I,67.609)

(319)
ʒe ne schulen, ic segge, makie na ma uuz of feste biheastes.
you-SUJET NEG FUTURE, I-SUJET say-PRES, make no more use-OBJET of feast promises-PL.

I say that you will not use the promises of feast anymore. (CMANCRIW,I.46.53)

(320)
... yet thar ye nat accomplice thilke ordinaunce but yow like.
... yet need you-SUJET NOT accomplice this decree unless you-OBJET like-PRES.

... yet, you do not need to enforce this decree, unless you want it. (CMCTMELI,220.C2.127)

then epistemic modals,

(321)
Acc žu shallt findenn žatt min word, ...
But you-SUJET FUTUR find that my word-SUJET, ...

But you shall find that my word, ... (CMORM,DED.L23.14)

[here the reading of SHALL is epistemic: I know that you will realize that...]

(322)
... michte for te serui že, wisdom for to queme že, luue ant wil to don hit; mihte, žet ich maʒe don, wisdom, žet ich cunne don, luue, žet ich wulle don al žet že is leouest.
... might FOR TO serve you-OBJET, wisdom FOR TO satisfy you-OBJET, love and will TO do it-OBJET; power, that I-SUJET can do, wisdom, that I-SUJET can do, love, that I-SUJET will do all-OBJET that this-OBJET is-PRES beloved.

... the might to serve you, the wisdom to satisfy you and the love and will to do it; the power I can use, the wisdom I can have, and love (which is everything) I want to give. (CMANCRIW,I.62.200)

(323)
The yeer of oure Lord 1391, the 12 day of March at midday, I wolde knowe the degre of the sonne.
The year of our Lord 1391, the 12 day of March at midday, I-SUJET CONDITIONAL know the degree of the sun.

In the year 1391 of our Lord, the twelfth day of March at midday, I woudl know the degree of the sun. (CMASTRO,669.C1.189)

(324)
I wolde seye, that he wolde geten hym sovereyn blisfulnesse;
I-SUJET CONDITIONNEL say, that he-SUJET CONDITIONNEL obtain him-OBJET soverein blissfulness-OBJET;

I would say that he would gain a soverein blissfulness; (CMBOETH,430.C1.82)

These examples allow us to show two things: 1) the process of grammaticalisation is set once and for all and 2) it is illustrated by the morphology of the modals which is to influence their meaning, that is past morphology but irrealis meaning (= conditional) (note: ).

Indeed, if we recall Structure (295) with the paradigm canst, we considered <cαn> to be its root which was in its turn not considered strictly speaking as a verb. From a semantic point of view, it does not imply the absence of meaning, since it means ``know, be able““. But if we consider the root <cαn> as a V-1, we could be led to think it has less meaning than V but not that it has no meaning at all. Nevertheless, by becoming a V, and then a time element after grammaticalisation, the meaning is to stabilize (see Stévanovich (2000) for the change between OE an ME).

We shall add some more examples of epistemic modals also taken from Denison (1993): 298-303. Examples (325) to (330) display a visible subject, Examples (331) to (339) do not.

(325)

sone hit męi ilimpen.

Soon it may happen. (a1225(?a1200) Lay.Brut 2250)

(326)

And if žou wynus it mai not be Behald že sune, and žou mai se.

And if you think it may not be behold the sun and you may see. (a1400(a1325) Cursor 289)

(327)
Vr neghburs mai žam on vs wreke.
Our neighbours may themselves on us avenge.

Our neighbours may avenge themselves on us. (a1400(a1325) Cursor 11963)

(328)

Sen žou ert both ong and fayre, žou mai haue childer to be žine aire.

since you are both young and fair you may have children to be your heir(s). (a1425(?a1350) 7 Sages(2) 2843)

(329)

žough že fflame of the ffyre of love may not breke out so žat it may be seyn, ... (1472 Stonor 123 I 126.36)

(330)
Že enbatelynge aboute... mai well be her feyned holynesse wherbi žei colouren al ere euele.
The battlements around... may well be their feigned holiness by means of which they disguise all their evil.

the battlements around ... may well be their feigned holiness by which they disguise all their evil. (?c1425(?c1400) Loll.Serm. 1.168)

(331)
grisen him mahte žet sehe hu...
feel horror him-OBL might that saw how...

He who saw how ... might feel horror. (c1225(?c1200) St.Juliana(Bod) 51.551)

(332)
ne schal him žurste neuere.
NEG shall him-OBL thirst never.

it shall never appease his thirst. (a1300 Žo ihu crist 85.24)

(333)
Vs shal euer smerte.
us-OBL shall always feel pain.

we shall always feel pain. (a1300 Sayings St.Bede(Jes-O) 83.336)

(334)
hym shall not gayne.
him shall NOT gain.

it will do him no good. (a1500(?a1300) Bevis(Cmb) p.83, textual note to 1583-1596)

(335)
Ne žurhte že neuer rewe, myhtestu do že in his ylde.
NEG needed you-OBL never rue, might you put tourself in his protection.

You would never need regret putting yourself in his protection. (a1300 A Mayde Cristes(Jes-O) 96)

(336)
Mai fall sum gast awai him ledd, And es vnto že felles fledd.
may fall some spirit away him led, and is into the hills fled.

Maybe some spirit led him away and he(?) has fled into the hills. (a1400(a1325) Cursor 17553)

(337)
Hym thar not nede to turnen ofte.
him-OBL need NEG necessarily TO turn often.

He need not necessarily turn often. (c1450(1369) Chaucer BD 256)

(338)
Him may fulofte mysbefalle.
him-OBL may very often suffer misfortune.

he may very often suffer misfortune. ((a1393) Gower CA 1.457)

(339)
Hym wolde thynke it were a disparage To his estate.
him-OBL would seem it was a disgrace to his estate.

It would seem to him a disgrace to his estate. ((c1395) Chaucer CT.C1. IV.908)

These examples underline the fact that epistemic readings increase at the ME period.

These examples also stress out the setting of the grammaticalisation of modals. This process has a very strong influence on syntax: as we have already said, the syntactic positions which are higher than T are grammaticalised, unlike the positions above T. Modals are still generated under vModal but before Vocabulary Insertion, vModal fuses with T. In the previous chapter, we underlined that Mood was ``grammaticalised““; but, during the ME period vModal is to grammaticalise as well: deontic modals are grammatical items. And both epistemic and deontic modals are raising verbs.

3.6.2. Modality and syntax

Let us recall the structure of epistemic and deontic modals:

(340)
(340.a)
Tree 88 Tree 89

In Section 3.6.1, we stressed the existence of four functional heads Mod (according to Cinque: ModNecessity, ModPossibility, ModVolition and ModObligation. So there seems to exist functional heads for modals in ME whether they are epistemic or deontic. However, we shall not adopt Cinque“s multiplicity of functional heads, but two modal functional heads Mood and vModal. Both have the following features: [±deont], [±past], [±realis], but vModal will have a [+deont] feature and Mood a [-deont] one. The [±past] feature depends on the nature of the infinitive clause, that is whether it is a present or a past infinitive; the [±realis] is context-dependent. For each form, let us give a table where the italicized vowel represents the root and the bold morpheme the verbal (dental) inflection.

(341)
Infinitive [-past] [±irrealis] [+past] [±irrealis]
MOTEN mot moste
OWEN owe ahte
WILLEN wil wolde
CUNNEN can cuže
DURREN dar dorste
MOWEN mai mighte
SCULEN shal sholde
ŠURFEN žarf žorfte
UNNEN an uže
WITEN wot wiste

Let us take CUNNEN: with the form can, we know we have chosen a [-past] form (because of the root vowel <a>); with the form cuže, we know we have a past form thanks to the root vowel <u> and the dental suffix <že> indicating the [+past] tense. Both forms can be [±realis].

3.6.3. Modal functional heads and infinitive structures

In the previous chapter, we defined two functional heads corresponding to each reading – epistemic or deontic – that could be made of the preterite present verbs.

According to the ME examples we have just analyzed in Section 3.6.2, the head vModal governs deontic modals and the Mood epistemic ones. As far as negation is concerned, we still find ne which no longer merges onto modals, and we more and more find not.

(342)
& žach žurch me ne schulde hit neauer beon iupped.
& then because me NEG should it-SUJET never be open-P.PASSE.

and then, becuase of me, it should never be opened. (CMANCRIW,II.70.795)

(343)
Certes, quod she, the wordes of the phisiciens ne sholde nat han been understonden in thys wise.
Certainly, said-PRET she-SUJET, the words-SUJET PL of the physicians-PL NEG should NOT have been-P.PASSE understood-P.PASSE in this way.

Certainly, she said, the words of the physicians do not have to be understood this way. (CMCTMELI,226.C2.360)

(344)
And therfore seith Salomon, ``The wratthe of God ne wol