In Example (133)， the MV must co-occurs with the MSA really.Here， really indicates the intensity exceeding a standard， and prediction (modalization).Hence， the subjective use of must is mitigated by the objective use of yet.In Example (134)， the pattern of “may yet” appears to be used in two types of context： (a）when something concrete is being predicted in a fairly short though unspecified time frame； (b）when predictions are less tangible and basically unverifiable.Walsh (2004： 335-348）explores the roles of MVs in expressing prediction in financial news articles， and finds some typical patterns of the co-occurrences of MVs and adverbs， e.g.may yet， might yet， could yet， may well， and might well.Thus， it is reasonable to say that the MSA yet conveys a certain standard in many cases， given the genre and tenor involved.In Example (135)， actually as the MSA of the MV should demonstrates a personal view instead of a shared standard； this MSA is subjective and has a speech-functional role.
Groenendijk et al.(1996： 182-185）argue for a dynamic interpretation about modality in that meaning is information change potential， that there are two kinds of information-information about the world and discourse information， and that the two kinds of information are linked to each other.It is believed that mostly the dynamic nature of the combinations between MSAs and MVs makes the combined modal meanings varied， and that such co-occurrences relate to both the actual world and the discourse world.
4.2.1.4 Balance between modalization and modulation
In SFL， modal devices expressing probability & usuality are put in the category of modalization， whereas those expressing obligation and inclination in the category of modulation.Modalization and modulation are related to information， and goods-&-services respectively.It is hard to classify MSAs of the categories other than that of probability & usuality according to the system of modalization and modulation.Instead， it is postulated that they may have a sense of modalization or modulation depending on the contexts involved.Consider the following：
(136）Admittedly he might already have done so.
(137）…the referral procedure provided for in Article 9 should only be applied in exceptional cases…
In Examples (136)， there are two MSAs for the MV might， i.e.， admittedly (MSA of clarification）and already (MSA of limitation).As the MV might here expresses modalization (i.e.prediction)， these two MSAs further specify the modalization of the MV.In Example (137)， the MSA only conveys limitation for the scope of legal regulation application indicated by the MV should.As should in Example (137）expresses modulation (i.e.obligation)， the MSA only adds more modulation to the proposition involved.
This understanding is significant for people to realize the roles of MSAs as a means of modality supplementing.MSAs can be classified in the system of modalization and modulation， just like the MVs they co-occur with.They are used to balance the modalization or modulation involved in three ways， i.e.retaining， strengthening and weakening.
4.2.2 Modal synergy
MSAs can help MVs better convey attitudes or judgments.In terms of ranks and mood structure， MSAs help MVs play the role of negotiation seen from the perspective of clause as exchange.MSAs can also make MVs better express evaluative functions.Again， this is a reflection of modal synergy.The modal synergy achieved is related to the impacts of contextual factors such as cognition， pragmatics and genre.
4.2.2.1 Variations in ranks
As MSAs are parasitic on MVs， they form a rank， i.e.modal verb group (MVG).Hence， there are variations in ranks as long as MSAs co-occur with MVs.
In Example (138)， the adverbial group has only a head without modifier.In Example (139)， the adverbial group has a head with a pre-modifier.In Example (140)， the adverbial group has a head with 5 pre-modifiers.These examples indicate that adverbial groups could be long or short depending on the contexts involved.Yet， it is known that MSAs could be prepositional groups as well.Halliday & Matthiessen (2004： 359）state that prepositional group is different from prepositional phrase.The former is a modifier-head structure expanded from and functionally equivalent to a preposition； the latter is not an expansion of anything but a clause-like structure in which the process/predicator function is performed by a preposition and not by a verb.Consider the following：
It can be found that in prepositional groups head usually comes before modifier， whereas in adverbial groups (cf.Examples (139）and (140)）the situation is just opposite.The situation becomes more complex when modal adjuncts become MSAs.Consider the following：
In Example (143)， should (MV）and of course (a prepositional group and an MSA）have their own group structures.Yet， when they form a modality supplementing pattern， there is an overall structure of MVG.Similar situation happens in Example (144).In both situations， MSAs become modifiers and β， while MVs are heads and α.In this way， there are variations of ranks as far as MSAs are involved.Generally speaking， MSAs and MVs are in the same clauses or clause complexes.However， genre may have impacts on the positions of MSAs and MVs.It was mentioned earlier in Chapter 3 that MSAs and MVs may occur in different clause complexes in conversations.
Generally speaking， in terms of ranks there are three grammatical patterns of modality supplementing through MSAs：
A.Word (or word group）+ word (or word group)
B.Word (or word group）+ clause
C.Clause + clause
Consider the following：
(145）…whereas the relevant legislation must henceforth apply to all statistics relating to the trading of goods between Member States…