Difficulties in explaining the various uses of the moods especially occur in the case of the subjunctive, which, as we have seen, occurs in all three sentence types (declarative, interrogative, imperative). I will limit myself to the subjunctive.
In describing the use of the subjunctive, some linguists take as a starting-point the intepretations of the subjunctive in the contexts in which it occurs. More or less comparable interpretations are organized as much as possible and then labelled as a certain use. This approach resembles the treatment of the cases (see p. 63). A representative of this approach is K.–St. Contextual elements are frequently adduced as evidence, as we have seen above in crosssection 10.2.2. Besides constituents in the context, the translation to the native language of the linguist often plays an important role. Handford, for instance, supports his view that at one stage an element of the meaning of the subjunctive was that of `intention' with the statement that `there are other passages where the translation `I will' gives a perfectly good sense' (1946: 39). 
A second group of linguists explains the Latin uses as much as possible from the origins of the subjunctive. A representative of this approach is Sz. I first discuss one such diachronic explanation. Then I return to some synchronic views.
(i) The origin of the Latin subjunctive receives, and has received, much attention (see Sz. 329–30).  Scholars agree that the Latin subjunctive is used in situations in which, for example, Greek would use the optative and the subjunctive, while there are also many formal similarities. From this it is concluded that the Latin subjunctive is a merger of the Indo-European optative and subjunctive.  This leads Sz. to organize (with some reservations) his chapter on the subjunctive as follows:
(a) uses of the old optative: wish, concession, irrealis (330–5); (b) uses of the old subjunctive: jussive, prohibitive, deliberative (335–8). 
(a) uses of the old optative: wish, concession, irrealis (330–5);
(b) uses of the old subjunctive: jussive, prohibitive, deliberative (335–8). The difference in negation (ne/non), as we have already seen, does not correlate with this classification.
(ii) The views described hereafter do not involve diachronic considerations. K.–St. distinguish there uses of the subjunctive, viz. the `potential' subjunctive (I.176–80), the `volitive' subjunctive (I.180–95) and the `conditional' subjunctive (I. 195, limited to imperfect and pluperfect subjunctive forms). They argue that all these uses can be explained from a common denominator, viz. that all `present something as thought, as imagined'  (see Sz. 326: `Modus der Vorstellung'). Opposed to this are the indicative and the imperative as `Modus der Wirklichkeit' (`mood of reality') and `Modus des Befehls' (`mood of orders'), respectively. Others distinguish two uses, viz. the volitive subjunctive (negation ne) and the non-volitive subjunctive (negation non), for
which no common denominator can be found.  A third approach explicitly considers the various uses as contextually determined within a kind of continuum of meanings that cannot very well be captured by a common denominator.  The unity of the subjunctive consists in its being opposed to the indicative and the imperative. 68a
To some extent, my own view has already been expressed on p. 194. I observe that in declarative sentences there is a semantic opposition e.g. (non) facio/faciam/facerem. I have described this difference with the terms `factive', `possible', `purely hypothetical', respectively. We find the same difference in interrogative sentences. In imperative sentences there is no semantic opposition (ne) facit/faciat, but there is an opposition (ne) faciat/faceret. In this case there is a contrast between possible and hypothetical. This has been represented in table 10.4 (see also table 10.1 on p. 190.) 
|Sentence type||Indicative||Subjunctive I (present/perfect)||Subjunctive II (imperfect/pluperfect)|
(77) quonam haec omnia nisi ad suam perniciem pertinere? (`To what did all this lead, apart from his own downfall?', Caes. Civ. 1.9.4) [cf. quonam haec pertinent]
(78) quis hoc sibi persuaderet sine certa spe Ambiorigem ad eiusmodi consilium descendisse (`Who would make himself believe that A. had developed this plan without certain expectations?', Caes. Gal. 5.29.5) [cf. quis persuadeat]
(79) quid illum facturum fuisse si … adversa pugna evenisset (`What he would have done if the fight had finished badly', Liv. 8.31.5) [cf. quid fecisset]
Pinkster, Harm (1942-) , Latin Syntax and Semantics [info], xii, 320 p.: ill.; 24 cm. [word count] [Pinkster].