'Nufaire, bel frere, nufaire tel sotie' (Liv Reis 81)
24 "The maiden answered: 'Do not do, stepbrother, do not do such foolishness.'"
This nu may however result from the contraction of ne and le rather than be a version of non (David Trotter, personal communication).
Which distinguishes this construction from others that have also disappeared, mainly the focusing si X non (a.) and embedded uses in relatives (b.) (Buridant 2000: 702-5)
a. Mais n'ad talent li facet se bien nun (Roland, 3681; Reid 1939: 305)
"But he hasn't done his will if well not." i.e. 'He hasn't done it if he hasn't done it well', 'He has done it'
Je ne dois pas vous relater qui se comporta bien et qui se comporta mal (Buridant 2000: 703)
"I don't have to tell you who behaved well and who [did] not."
25 Unless referenced, the glosses and translations are mine.
26 This applies not only to root, but also to embedded clauses (see Chil Buriles disoit ke le terre ke Esclas tenoit devoit estre soi, et Esclas disoit ke nondevoit"Chil Buriles was saying that Esclas' land should be his, and Esclas was saying that it shouldn't", Reid 1939: 307). Such cases are counterexamples to the claim by Godard and Marandin (2006) that only root clauses can host activated propositions.
27 Preverbal non is thus different from the sentence-anaphoric use of non that still exists today; the latter clearly is context dependent given its anaphoric nature but does carry with it words used in the preceding context.
28 That the construction is used with a limited set of verbs is true, but apart from the fact that there is a range of tense and person conjugations involved that should put in doubt any claim that an entirely set phrase is involved, it is tangential to the fact that the context is activated; a dependence on the antecedent context is found in the sequences where preverbal non is found whether they are a set phrase or not.
29 Criterial tests are proposed for contemporary languages in Larrivée (under evaluation) that involve putting the sequence under consideration in characterised textual environments, but these suppose acceptability judgements that can hardly be provided for states of languages where no native speakers are available.
30 One of the uses of point is with a verb that has been used in the immediately antecedent context:
a. et tout soit ensi qe les blees dedeins meisme le maner cressantz sont enmuriz, et mestier fuist qo ascun se entremeit de les faire entrer, par doute de perier, vous ne vous ent(re)medlezpoint, com faire duissez, a ceo qe nous est fait entendont,
"and it was such that the wheat inside growing in the same way were walled, and the service was that someone should have intervened to get it in, lest it withered, you didn't intervene, as you should have done, according to what we were told"
but this is not the case of the other:
b. Mes nous lui commaundames qil savisast soulement ovesque vous quoi seroit a faire de cel endroit. Parquoy nous nous agreomspoint ceo que il ent ad fait.
"But we asked him whether to see with you what should be done in that place. Thus we do not agree with what he has done with it."
31 The first example further contains a non-sentential contrastive construction nonpas, which constitutes accommodated activation. It recurs in two other occasions in the corpus:
a. Kar a la corone apartent nun pas soulement crueaute e reddur de justice, mes plus pite e misericorde.
"Because to the crown belongs not only cruelty and severity in justice, but even more pity and mercy."
b. Et auxi en voz dites lettres feut contenuz, qe, par procurement des uns de noz com-moignes, ses enemys com est dit, par articles ne pas veritables et malement compassez, et collusions par eaux entremises, est atort grevez,
"And also in those letters of yours it was suggested that, by the persuasions of one of our monks, his ennemies as is said, by articles not truthful and badly constructed, and collusion between them, you unfairly wrong him"
Note that the non-verbal context does not forbid ne, which might be required to license pas – and in other cases in the same corpus mie. One sentential contrastive case is attested.
c. pur co ke il nest pas en sa garde, einz e la nostre.
The contrast can reasonably be argued to force accommodation of activation: the (c) example does imply that one could believe the person in question to be in the care of a third party.
32 One reviewer raises the possibility that children’s perception of formality markers might differ from that of the adult community. Study of the adult’s production and perception would be useful in apportioning this.
33 Even lower frequency is found in the socially representative corpus of Quebec French by Sankoff and Vincent (1977).
34 The new online corpora of spontaneous speech Discours sur la ville – Corpus de Français Parlé Parisien des années 2000 (CFPP2000) reveals a similar situation to that in Lancom.
35 As pointed out by a reviewer, this looks more like a linear process than a cycle. For a full assessment of the notion of cycle for French negation, see Larrivée (2009).
36 An illustration of this mechanism relating to variation is documented by Hack (2009), who shows that in the Dolomitic Ladin dialects where particle pa becomes a compulsory total question marker, it has lost all the 'emphatic' values it has in the varieties where it is not required.
37 Among these pragmatic values is the denial of a previous assertion represented by activation. Denial of a previous assertion is often identified as the basic function of all negatives in the research tradition, according to one reviewer who points to Givón, Horn and Miestamo. It is true that Givón and Horn discuss the informative asymmetry of negation, but they do not to my knowledge assert that all negatives represent denial. If denial were the function of all negatives, then clearly that would make the notion of activated negation redundant. The way in which this notion is distinctive is that it supposes material accessible to the hearer rather than just to the speaker, and that in this sense, sequences like She wasn't moving, she wasn't breathing, we couldn't do anything are in most cases not normally activated while being certainly negative.
38 One reviewer notes that "presumably there are some cases (e.g. morphological regularisation) where (in)frequency drives (rather than follows) change." This conclusion could only be supported by extensive quantitative data for which the possibility of any status change would be excluded, certainly a desirable endeavour for future research.
39 One reviewer notes that the "explanation by prescriptive teaching is not sufficient: expletive negatives were in any time criticized by normative grammarians, and nevertheless are always in use, sometimes against normative grammar." Indeed, it is not teaching itself, but rather the specialised contribution of the item reinforced by teaching that ensures its longevity.