Modality as part of language is an extremely complex phenomenon whose definition and description have been one of the most persistent problems for linguists. The concept of modality in ordinary language includes attitudes which are similar to, but not identical with the logical, epistemic and deontic modalities. It includes feelings and emotions as well as other concepts that do not exist in modal logic. Therefore, natural language modality should not be confused with logical modality. 

        The complexity of this phenomenon is so great that there is little agreement in the literature as to how modality should be defined. But we all agree that “modality expresses the attitude of the speaker towards what he is saying.”(T.Duţescu and Coliban, Gramatical Categories of English, p.275). In other words, as Loreto Todd and Ian Hancock say, “modality refers to the attitudes expressed by a speaker towards the statement or proposition being made.” (Loreto Todd / Ian Hancock,International English Usage, p.22). 

       They also add that such attitudes may express ability, compulsion, desire, insistence, obligation, permission, possibility, willingness and uncertainty.

      Certain grammar postulate a modality component in their organization. In Filmore’s Case Grammar , for example, every sentence is decomposed into a modality constituent and a preposition, that is a set of relationship involving verb phrases and noun phrases, not different in essence from the logical notion of proposition.

       What is important in this view point is the idea of separating modal meaning from the basic meaning. Although modality is often characterized as a verbal category, since it is at the level of the verb phrase, modality applies to the whole statement. So, it should be taken as a sentence category.

     It is relevant to say that modality can be seen semantic point of view. Wilfried Rathay tries to understand whether modality constitutes a grammatical category, i.e. a (morpho-syntactic category, or whether it should be looked upon as a semantic phenomenon. He considers that modality has to be regarded as one of the main features of sentence or, as some linguists maintain, as an obligatory characteristic of every sentence. (cf. Gordon /Krilova, Pantilov, Mühlner). There are, however, divergent opinions about what be included under this term. In some cases only the general feature is taken to be essential: “ the relationship between the statement affirmed in the sentence and reality as established by the speaker.” Other linguists emphasize the twofold aspect of the notion “ modality” and differentiate between: a) “the relationship of the content of the sentence to reality (as regarded by the speaker)” and b) ”the attitude of the speaker towards the content of the sentence.” (cf. Achmanova). Sometimes a third aspect is mentioned which is supposed to specify “the relationship of the subject of an action, process or state towards this action, process or state.” (modality expressed by modal verbs: He can swim , He may come , He must do it. ) (cf. Sommerfeldt, Schmidt, Zolotova ).

    The first aspect is quite often referred to as “objective or basic modality and it is expressed by the morphological category “mood”. (indicative, subjunctive, imperative”). So, the main distinction should be drawn between “real” and “unreal” events.

       The second aspect is usually called “subjective modality”. It permits the speaker to evaluate the content of a sentence in a particular way. In contrast to objective modality, it is regarded as optional, i.e. a sentence may or may not contain an element of objective modality.

        Halliday reduces the term “modality” to “the speaker’s assessment of the probability of what he is saying or the extent to which he regards it as self-evident.” The meanings of the modals such as permission, ability, obligation, compulsion etc. are considered to belong to a different system, which he calls “modulation”.

      Sometimes modality is identified with the “communicative intention” of the speaker, i.e. modal types are set up according to the main types of sentences with respect to their function in a particular communication situation. (assertive, interrogative and imperative modality). This view is rejected by many linguists who maintain that the communicative intention of the speaker towards the interlocutor is something different from the relationship between the content of a sentence and reality.

        I would add here Hemeren’s point of view. He distinguishes three main types of modality: internal, external and neutral.

(a) In the “internal modality” the constituent expressing modality and the following verb share the same subject. Consequently the relationship may be illustrated thus: 

John able (John go). 

In the surface structure this may be realized as, for instance: 

John can go.    or

John’s ability to go.

(b) In the “external modality” the constituent expressing modality and the following verb have different subjects, as in:

Dick permit John (John go).

This can be realized as:

“You may go,” said Dick to John.  

(c) Finally, in the “neutral modality” the subject of the constituent expressing modality is the clause consisting of the following verb and its subject, as in:

Probable (John go).  

where John goes is the subject of probable.

This can be realized as:

That John will go is probable.

or

The probability of John going.

        Most of the traditional grammars do not give a coherent presentation of the category of modality. Modal types and meanings are usually described isolated from each other according to the grammatical categories or parts of speech they belong to. (mood, modal verbs, modal nouns etc.) The way in which transformational generative grammars have so far dealt with questions of modality seems even less satisfactory. Earlier versions of TG enumerate some modals as elements of the auxiliary constituent derived from.

       Others, such as Roberts or Catell say a little more about the modals, but do not touch upon any fundamental questions. Others either deal with questions concerning the classification of the modals or try to present a few semantic features for certain modal verbs.

        The necessity of resorting to the notion of “speaker” in the definition of modality situates its study within the domain of pragmatics. While the formalized logical structures are rigid and limitative, in natural languages modality can be expressed by a great variety of means. The speaker may express his attitude in many ways, some of which are linguistic and others extra-linguistic.

        The linguistic phenomena within the range of modality are numerous and diverse, and can be found at all levels of language analysis.

(a) At the phonological level we have the phonological means such as intonation and accent.

By varying our intonation, we can modify meaning as in:

You’ll go all right! =  I’ll see to it!

              Over my dead body!

What is also important is the relation between modality and accent:

She is lovely! (“Vai, ce draguta e!”)

as well as the relation between modality and sound length:

Wha-a-a-a-t? (“Ce-e-e?”,”Cu-u-u-u-m?”)

(b) At the lexical level, adjectives like necessary, possible, desirable, recommendable contain elements of modality, as well as adverbials like perhaps, maybe, nouns like necessity, probability, and verbs which have semantic features of modality like doubt, think, wish.

        Quirk (Quirk, 1972,p.512) says that there is a group of disjuncts that convey some attitude towards what is said. Only the most common adverbs are included. Those with a participle stem in –ing are the most productive class of adverbs as attitudinal disjuncts. He classifies them into two groups:

i. Disjuncts that do not normally express the view that the speaker’s judgement applies to the subject of the clause to which the judgement is attached.

• What is said is judged to be strange or unexpected: amazingly, astonishingly, curiously, funnily enough, incredibly, ironically, oddly, remarkably, strangely, suspiciously, unexpectedly.

• What is said is judged to cause satisfaction or the reverse: annoyingly, delightfully, disappointingly, disturbingly, luckily, sadly, tragically.

• Other judgements: amusingly, hopefully, mercifully, preferably, significantly, thankfully. 

ii. Disjuncts that express a judgement on what is being said as a whole and normally apply the same judgement simultaneously to the subject of the clause.

• Judgement is passed on the wisdom or skill of what is described: artfully, cleverly, cunningly, foolishly, prudently, reasonably, unreasonably, sensibly, shrewdly, wisely, unwisely.

• Judgement is passed on whether what is said is right or wrong: correctly, incorrectly, justly, unjustly, rightly, wrongly.

        Quirk also includes here some adverbial structures with a similar semantic force: to my regret, to our surprise, with justice, of course, no doubt, what is remarkably etc.

To my regret, he did not accept our offer.

They arrived, to our surprise, before we did.

With justice, they have assumed their place the nation’s heroes

To the great admiration of all the people, he plunged into the water and rescued the child.

Of course, nobody images that he will repay what he borrowed.

It was no doubt that she would buy that dress.

What is even more remarkably, he manages to inspire confidence in the most suspicious people.

(Quirk, 1972, p.515).

     Halliday (Halliday 1994, p.49) speaks about modal adjuncts. They show the speaker’s judgement regarding the relevance of the message. So, he uses another term for Quirk’s disjuncts. Among the principal modal adjuncts there are: probably, possibly, certainly, perhaps, maybe, usually, always, sometimes, (n)ever, often, seldom, of course, obviously, clearly, surely, in my opinion, personally, to my mind, frankly, to be houest, to tell you the truth, honestly, really, believe me, seriously, evidently, apparently, no doubt, (un) fortunately, to my delight,  regrettably, hopefully, broadly speaking, in general, on the whole, strictly speaking, in principle, foolishly, wisely, to my surprise, surprisingly.

(c) At the morpho-syntactic level, we have moods and modal verbs as well as the highly specialized modal auxiliaries can, may, etc.; different types of sentences of word order (sentence structure) may also mark modality.

• If we take into consideration word order, the clearest examples are the differences between statements, questions, orders and blessings curses:

 You are going away.

Are you going away?

Go away!

May you never know rest!

• There are also types of subordinate clauses (concessive and conditional clauses) that are incorporate in them a certain attitude of the speaker R.W.Zandwort (Zandwort 1958, 123) says that modality characterizes the fundamental classification of the sentence:

(1) Declarative sentences(affirmative or negative, expressing different modalities)

(2) Interrogative sentences (they express interest, curiosity, wonder, perplexity etc.)

(3) Exclamatory sentences (they express admiration, approval, disapproval, disapproval, astonishment etc.)

(4) Imperative sentences (which express an order, advice, a request etc.)

(d) If the problem of modality is approached in a semantic perspective, all language levels should be considered globally.

      So, as Wilfied Rathay says, linguistic modality should be looked upon us “a semantic category which modifies a language utterance in a particular way and which may be expressed by number of different means in a given language (mood, modal verbs, modal tense and aspect forms, lexical verbs, word order, intonation etc.)” (Rathay 1979, 105). And he continues that “one and the same modal type can be expressed by different means in a particular language.”

He may/might be there.

It is possible that he is there.

I think/consider it is possible that he is there.

Perhaps/Maybe he is there.

        Many classifications are possible at the semantic level, but must grammarians distinguish between the epistemic and deontic modals.

        Grammarians have noticed that the modal verbs are of two different natures: (a) one of a concrete nature being called deontic abstract nature expressing permission, volition, obligation, and (b) a more abstract nature expressing epistemic values (from possible, through probable, up to certain). The first nature of the modal verbs is usually referred to in the literature as the root (or deontic) sense, while the second nature is referred to as the epistemic sense.

        Hofmann (Hofmann 1993, 106) describes the deontic and epistemic modals starting from the idea that “we also use modals commonly to describe what is expected or demanded of us. These are deontic modals. The following examples depict what the speaker or someone else makes impossible, possible or necessary for the subject.

 You may come now.

You can’t come now.

You shall come now.

      While the deontic modals have a subject orientation, describing what is necessary, possible or impossible for the subject of the sentence, the epistemic modals can be seen as logic-oriented (sometimes called “alethic”). These epistemic modals seem to be totally devoted to the end, often when the other proposition is expressed as an independent sentence.         In the example below the tow sentences express the idea that the modalized sentence is a necessary conclusion from the other, whatever is the order of the sentence.

John’s umbrella is wet. It must be raining.

It must be raining. John’s umbrella is wet.

Similarly, indicating a conclusion of possibility and conclusion of impossibility.

It must be raining. It’s awfully dark out.

It can’t be raining. There’s not a cloud in the sky.