(259）The constitution, it is true, secured every essential guarantee to the institution while it should last, and hence no argument can be justly urged against the constitutional guarantees thus secured, because of the common sentiment of the day.Those ideas, however, were fundamentally wrong.They rested upon the assumption of the equality of races.This was an error.
(260）Many who hear me, perhaps, can recollect well, that this truth was not generally admitted, even within their day.
(261）The reply I made to him was, that upon his own grounds, we should, ultimately, succeed, and that he and his associates, in this crusade against our institutions, would ultimately fail.
Examples (259）to (261）are from the same public speech.In these three examples, three MVs are employed: should, can and would.In Example (259), should is used for modulation, and can for modalization.The MSAs hence and justly supplement the modalization of can from the categories of reasoning and evidence respectively.In Example (260), the MV can may be used either for modulation or modalization if the MSA perhaps (mood adjunct of probability）does not occur.However, with the occurrence of the MSA, the MV can demonstrates modalization.In Example (261), the MVs should and would denote modulation and modalization respectively, and the MSA ultimately (mood adjunct of reasoning）occurs twice to supplement the two different kinds of modality.
Hence, in some cases there is a balance between modalization and modulation by means of MSAs in PS.This shows the stance of negotiations and a balance between solidarity and deference in this genre.
The situation is similar in JI.Consider the following:
(262）I think that’s about it right now that I can talk about.There’s other stuff that sort of comes up, but I can’t really talk about it.
(263）Well, it is difficult to switch back and forth between being an actor and a director, because as an actor, you show up for your little bit and they can usually move you around to different places, and you show up usually for a maximum of four hours, which is the longest a recording session would be, and even if they need to split it up, maybe they can fit you into the schedule.
Examples (262）and (263）are from the same journalistic interview.In Example (262), the MV can and its negated form can’t both convey modulation, and the MSA really supplements modulation from the category of confidence.In Example (263), the MVs can and would are used for modalization, and the MSAs usually and maybe supplement modalization from the categories of probability and usuality.
Thus, in many cases, there is a balance between modalization and modulation by means of MSAs in JI.The stance of negotiations and a balance between solidarity and deference are obvious in this genre.
Similar situation occurs in AW.Nevertheless, some interesting features are found by comparing the attractions between MSAs and MVs in the four genres:
A.The MVs should, must and shall can be used for both modulation and modalization, whereas the MVs can, could, will, would, may and might are used for modalization only on many occasions in PS, JI and AW.Thus, the MSAs co-occurring with these MVs show different functions of modality in these genres depending on the contexts involved.
B.Almost all the MVs are used for modulation in LR.Hence, the MSAs co-occurring with MVs normally supplement modulation in this genre.
C.In general, modalization is more frequent in AW than in PS and JI.Therefore, MSAs supplement modalization more than modulation in AW than in PS and JI.
Consider the following:
(264）It can be assumed that the silences that characterize newspaper articles differ, at least slightly, from those of the specific part-genres of written scientific discourse.
(265）Distressingly, these authors conclude that social scientists should adopt conventional methods of mainstream science.
Examples (264）and (265）are taken from AW.In Example (264), the MSA at least supplements the modalization of can from the category of clarification; in Example (265), the MSA distressingly supplements the modulation of should from the category of confidence.These two examples dovetail with the conclusion that the MVs should, must and shall can be used both deontically and epistemically, whereas the MVs can, could, will, would, may and might are used epistemically only on many occasions in AW.
Therefore, it can be concluded that MSAs function in a way similar to MVs in terms of stance and social distance, based on generic nature and communicative needs.
6.7 Occurrences of MSAs in types
It is found that MSAs differ in distributions according to types in the four genres.Table 6.7 indicates the distributions in this regard.
The findings shown in Table 6.7 are significant in revealing the stance and social distance involved.
First, in PS (158 occurrences）and JI (197 occurrences), speakers are very careful about limiting what they say, and AW (101 occurrences）also displays such a tendency though it is not so strong.In contrast, in LR (45 occurrences), the MSAs of limitation scarcely occur.This is most likely related to the exactness of wordings in this genre.Legislators know that they cannot speak so tentatively as in other genres like JI or PS.
Second, the MSAs of probability & usuality in JI (231 occurrences）outnumber those in PS (171 occurrences), AW (130 occurrences）and LR (120 occurrences).This may have much to do with the strong colloquial features of JI.In face-to-face talks, prediction is common; hence, the MSAs expressing likelihood and habits are not rare.
Third, the MSAs of confidence occur most often in JI (233 occurrences).As is known, confidence is much related to casual talks, where discourse participants are not constrained by social status or power.Therefore, JI as a genre showing the close relationships between interviewers and interviewees has the biggest number of this type of MSA.