LINGUIST List 12.2018

Fri Aug 10 2001

Review: Palmer, Mood and Modality, 2nd ed.

Editor for this issue: Terence Langendoen <>

What follows is another discussion note contributed to our Book Discussion Forum. We expect these discussions to be informal and interactive; and the author of the book discussed is cordially invited to join in.

If you are interested in leading a book discussion, look for books announced on LINGUIST as "available for discussion." (This means that the publisher has sent us a review copy.) Then contact Simin Karimi at or Terry Langendoen at


  • Sharbani Banerji, review of Palmer, Mood and Modality, 2nd ed.

    Message 1: review of Palmer, Mood and Modality, 2nd ed.

    Date: Thu, 9 Aug 2001 22: 55: 27 +0530
    From: Sharbani Banerji < net">sharbevsnl. net>
    Subject: review of Palmer, Mood and Modality, 2nd ed.

    Palmer, Frank R. (2001) Mood and Modality, 2nd ed. Cambridge University Press, hardback ISBN: 0-521-80035-8, xxi+236pp, $64. 95, Cambridge Textbooks in Linguistics (1st ed. 1986; paperback ISBN: 0-521-80479-5).

    Sharbani Banerji, Centre for Applied Linguistics and Translation Studies, University Of Hyderabad, India.

    This book is a typological study of Mood and Modality, and is the second edition of the original volume, which was published in 1986. Frank Palmer points out that the interest on Mood and modality, as well as on grammatical typology in general, is quite a recent affair, and has grown mostly after the publication of the first edition of the book.

    This is confirmed by the fact that whereas the symposium of Mood and modality held in the University of New Mexico in 1992 brought together over forty researchers, and resulted in the publication of eighteen papers (Bybee and Fleischmann 1995), yet, in contrast, a workshop on modality at the International Congress of Linguistics only ten years before had attracted only four scholars.

    Modality is concerned with the status of the proposition that describes the event, and is a cross-language grammatical category that can be subject to typological study. Drawing upon data from a large number of languages, Palmer proceeds very systematically in the typological study of modality, proving at the very outset that grammatical typology cannot be undertaken on the basis of purely formal grammatical marking. It is thus done on the basis of 'notional' criteria. The term 'notional' is used by Palmer to avoid the debate about semantic vs.. pragmatic etc. Thus, this kind of work requires that various kinds of modality be first identified, and then it needs to be checked how they are represented in various languages. Though languages vary in the categories that are treated as Realis or Irrealis, there is a lot of similarity too, amongst widely different languages. Lot of questions come to mind, which perhaps can be answered only when more detailed work on 'Modality' is done on individual languages, so that relationship of modality with other modules of grammar can be better understood.

    Thus, this work is useful for students, research scholars as well as for professionals working on 'Modality', be in an individual language, or as a typological study. The detailed classification of modality that has been undertaken in this work, in fact, without which the typological study wouldn't have been possible at all, is the most valuable and the most essential guide to anybody who is working on modality. There is a lot to learn about the methods of 'typological study' from this book.

    After the introductory chapter, the book is basically divided into two parts. The first part, chapters 2, 3 & 4 is concerned with modal systems. Chapter 2 deals with propositional modality, with main sections on epistemic and evidential modality. Chapter 3 deals with event modality, with main sections on deontic and dynamic modality. Issues affecting both types of modality are considered in Chapter 4. The second part, chapters 5, 6 & 7 is concerned with mood. Chapter 5 deals with indicative and subjunctive and chapter 6 with realis and irrealis, while issues affecting them both are considered in chapter 7. Chapter 8 deals with the issue of past tense being used as a modal.


    INTRODUCTION: In the introduction, basic concepts are introduced. The detailed analysis of each of the categories of 'Modality' are taken up in the subsequent chapters. The difference between Realis and Irrealis, which is fundamental to the concept of modality, is defined. For example, whereas Realis portrays situations as actualized, knowable through direct perception, the irrealis portrays situations as purely within the realm of thought, knowable through imagination. However, languages vary in the categories that are treated as Realis and Irrealis; e.g., one language may mark commands as Irrealis and another may mark them as Realis, while another may not treat them as a system of modality at all. There are similar variations in the treatment of futures, questions, denials, reports etc.

    Many European languages use Indicative and subjunctive to distinguish between Realis and Irrealis. However many authors treat the distinction between indicative and subjunctive in terms of 'assertion' and 'non- assertion'; e.g., Lunn 1995, suggests that a proposition may be subjunctive for three reasons:

    1) The speaker has doubts about it's veracity

    2) The proposition is unrealized

    3) The proposition is presupposed.

    Mood and Modal systems:

    There are two ways in which languages deal grammatically with an overall category of modality:

    i) modal system

    ii) mood

    Both may occur within a single language, e.g., in German, and in Central Pomo. In most languages however, only one of these devices seems to occur or one device is much more salient than the other.

    Typically with mood, all or most clauses are either Realis or Irrealis: the system is basically ('prototypically') binary. The indicative marks clauses as Realis and subjunctive as Irrealis. In native American languages and languages of Papua New Guinea, the distinction has been made in terms of grammatical categories realis and irrealis, rather than in terms of indicative and subjunctive. Strictly, there is no typological difference between indicative/subjunctive and realis/irrealis. Both make express the distinction between notional features of realis and irrealis and can be seen as an instance of the typological categories of Realis and Irrealis. However, there are some differences in their distribution and syntactic functions. E.g.,

    i) The subjunctive is found mainly in subordinate clauses

    ii) Irrealis and realis often cooccur with other grammatical markers

    iii) Both are often notionally redundant, but in different ways-- irrealis because of its concurrence with other markers, subjunctive, because, in subordinate clauses, its occurrence is determined by the type of complementizer.

    iv) Unlike the indicative/subjunctive system, realis/irrealis systems do not usually occur together with tense systems. In general, past and present are marked as realis, future as irrealis.

    Palmer admits that it is not always possible to draw a clear distinction between mood and modal system, since, in some languages, the overall system of modality has characteristics of both.

    The Classification of Modality in Modal Systems:

    Modality can be classified as Propositional Modality and Event Modality. Propositional Modality can be of two types:

    i) Epistemic

    ii) Evidential:

    With epistemic modality the speakers express their judgements about the factual status of the proposition, whereas with evidential modality they indicate the evidence they have for its factual status. Thus, epistemic modality can be: Speculative, Deductive or Assumptive. Evidential categories can be: Sensory and Reported.

    Event Modality in turn can be of two types: Deontic and Dynamic: With Deontic modality, the conditioning factors are external to the relevant individual, whereas with Dynamic modality they are internal. Thus Deontic modality can be related to obligation or permission, emanating from an external source, whereas Dynamic modality relates to ability or willingness, which comes from the person concerned.

    Thus the typological categories of Deontic are: Permissive & Obligative. Commissive too is a type of Deontic modality. It is defined by Searle as 'where we commit ourselves to do things', and is exemplified by the English 'shall'.

    The Dynamic modality are: Abilitive and Volitive.

    Other Types of Modality: Presupposed propositions are treated as irrealis, and many languages use subjunctive in the subordinate clauses. Yet, examples of presupposed propositions being marked as irrealis can be found in main clauses too, e.g., in Italian. Negatives and Interrogatives being 'non-assertive', sometimes occur within the modal systems or are marked as irrealis where there is marking of mood. In European languages, though in main clauses they are never treated as irrealis, there is evidence of such treatment in subordinate clauses. Wishes and Fears express attitudes towards propositions whose factual status is not known, or propositions that relate to unrealized events. They are thus partly deontic and partly epistemic. Many languages mark them as subjunctive in subordinate clauses, and some even in main clauses, e.g., Latin and Classical Greek. Past tense can be used to express 'unreality', 'tentativeness', 'potentiality' etc. This often functions independently of 'mood', but may interact with it. Past tense can also be used to express 'unreal conditional'. This use of past tense is referred to as 'modal-past'. Other important categories that may be marked as irrealis but are mostly found with mood are: Future, Negative, Interrogative, Imperative-Jussive, Presupposed, Conditional, Purposive and Resultative. Wishes (desiderative) and fears (timitive), and less commonly habitual-past.

    Complex Systems:

    Ideally, there would be discreet systems of modal markers, such that each of the typologically relevant modal systems was in a one-to-one correspondence with a formal system in each language. In practice however, the systems are not so orderly and simple. e.g.,

    i) Two different systems may use the same (or most of the same set of markers). E.g., in English, 'can' may be used as an epistemic, deontic or dynamic.

    ii) In few cases notionally identical modal categories may be treated as members of one system in one language and of another system in another language. e.g., 'Deductive', may be a term in a judgement system or in an evidential system as in English and Central Pomo respectively.

    iii) A single formal system often contains forms that belong to two or more modal systems. Thus in the German system of modal verbs, not only do most of the modal verbs function as both epistemic and deontic, but there are forms (sollen and wollen) that are clearly evidentials.

    iv) There may not be a simple one-to-one relation between the terms of a formal system and some of the typologically relevant notional features.

    v) The same formal system may contain not only modal categories, but others such as tense etc. In Ngiyambaa, there is a set of clitics that relate to features that are clearly modal, and to others that are not.

    Grammatical Markers:

    The grammatical markers of modality are very varied. Basically, there are three types of markers: i) Individual suffixes, clitics and particles; ii) Inflection; iii) Modal verb.



    This chapter deals with Propositional Modality in various languages. Propositional Modality are of two types: Epistemic modality and evidential modality. Epistemic Modality are of three types: Speculative, Deductive, Assumptive.

    English is one of the few languages which can express all the three notions using the three modal verbs. E.g.,

    i) John may be in his office (Speculative)

    ii) John must be in his office (Deductive)

    iii) John'll be in his office (Assumptive)

    Similar contrast is found in Danish (Germanic), Italian (Romance), modern Greek etc. It is also found in Tamil, in the form of verbal suffixes. They are also found in Ngiyambaa, Imbabura, Inga, a variety of Quechua, with finer shades of meaning.

    The contrast between Deductive and Assumptive are found in a number of other languages, but in most of them, Deductive and Assumptive occur together in systems that include evidential markers of report and sensation. e.g., Tuyuca, where the, markers correspond to the typological categories Visual, Auditory, Deductive, Reported, Assumptive etc. Similar contrasts are found in Wintu (N. California, now almost extinct), and other languages, once again, with finer shades of meaning.

    Presupposed modality:

    In English, 'May' can be used as a concessive, i.e.,, in the sense of 'although'. e.g.,

    He may be rich, but he is not very lucky.

    He may have been rich, but he wasn't very lucky.

    Similarly, German uses 'MOGEN'. Since with a concessive clause, the speaker does not indicate doubt about the proposition, but rather accepts it as true, in order to contrast one state of affairs with another, 'May' is hence used here not as Speculative, but as 'Presupposed'. In Latin, Italian and Spanish, a subjunctive will be used.

    Evidential modality

    Evidential categories are basically of two types, viz.,

    Reported and Sensory. Ngiyambaa has just two 'evidence clitics' (sensory and linguistic), but Central Pomo and Tuyuca have extended evidential system, e.g., 'visual', 'non-visual', 'apparent', 'second hand' and 'assumed'. Similarly, Hidatsa (Siouan, USA) has mood morphemes at the clause final position expressing various shades of extensive evidential system. Extensive evidential systems are found in Native American languages, and some in Papuan languages. Similarly in Ladakhi (Tibeto-Burman).


    Examples of reported (e.g., 'second-hand', 'linguistic evidence', 'hearsay') etc are found in Tuyuca & Ngiyambaa & Fasu. Reported is also found in languages that have a system of mood, but only where the system is joint, i.e.,, where the grammatical markers of Realis and Irrealis co-occur with

    Grammatical markers of other categories. This suggests that these categories are notionally 'realis'/'irrealis'. Thus 'reported' is marked as 'irrealis' in Hixkaryana. Similar forms are found in languages with predominantly judgmental systems, e.g., German and Danish, etc.

    There can be subcategories of 'Reported': e.g., 'second hand evidence', 'third hand evidence', 'evidence from folklore' etc. In some languages like Lega, a distinction is made between a report that is reliable and a report that is unreliable.


    Just as in case of reported, whereas some languages have a single category of Sensory, others have subcategories, e.g., visual and auditory. Ngiyambaa has a clitic which is used for sensory evidence involving all senses. Tuyuca has one marker for sight (visual) and another for all other senses (non-visual). Central Pomo has separate markers for both visual and auditory; similarly in the Papuan language Fasu.

    There are no attested examples of a specific marker for the senses other than seeing and hearing.

    Direct and indirect evidence

    Some languages may have a single marker of modality to indicate either what is said (report) or what may be inferred (deductive). e.g., Turkish, Sherpa, Abkhaz (N.W Caucasus) etc.


    Palmer notes that Oswalt (1986: 43) says of evidentials in Kashaya that they lie on a hierarchy:

    performative > factual > visual > auditory > inferential > quotative = Performative > Declarative > Visual > Auditory > Deductive > Reported.

    Those that precede have priority over those that follow. Oswalt says that the order seems to be universal, even for English, where evidential concepts are expressed by verbs.

    Interrogative and Negative are sometimes, found as members of an epistemic modal system. e.g., in Menomini (Algonquian USA) interrogative is a term in a modal system. Some languages use the same marker for both Negative and interrogative, since these can be seen as non-assertive. e.g., in Imbabura and Tiwi (Australia).

    Discourse and participants

    Modals have an important part to play in discourse, as the participants express their opinions and attitudes and, in general interact with one another. Cashibo has a full discourse system, and in particular forms for making statements, asking questions, giving replies and giving emphatic replies.

    In Khezha (Tibeto-Burman-Kapfo, quoted by Bhat 1999: 80-1), there are eight type of markers for Yes-No questions and eight for wh-questions, the choice being largely related to the features of discourse. In some languages there are systems indicating more direct reference to the participants in the discourse.


    It is generally the case that in languages with epistemic modal systems, there is a form that is unmarked for modality, and which, notionally, simply makes an unqualified assertion. It can thus be seen as Realis, and the modal forms as Irrealis. This form is identified as the Declarative. Thus for English, the declarative is marked by the absence of a modal verb. Since the declarative is unmarked for modality, it is not 'stronger' than a modal form. It simply asserts without indicating the reasons for that assertion or the speaker's commitment to it. It is perfectly possible for a speaker to say 'John is in his office' without being wholly sure or when there is strong evidence for it.

    Again, it is perfectly possible to express modal notions without the use of modal systems. This can be achieved through the use of lexical verbs as in:

    I think that Mary is in her office.

    I saw John in his office this morning.

    Grammatically, neither the first of these is an example of the modal judgement Speculative, nor the second is an example of the Visual evidential. They are both declaratives and the relevant notional interpretation is in terms of asserting that 'I think' and that 'I saw'.

    In terms of Realis vs. Irrealis, the declarative with no modal will be a Realis, whereas the modal forms will be Irrealis. However, this simple situation is not typical of all languages with modal systems. There are several different situations.

    In a few languages, the declarative is a term in the same formal system as the modal categories. e.g., in Ngiyambaa (New South Wales, Australia). Here past and present (declarative) occur in the same system as markers of Speculative ('irrealis'), and Deductive ('purposive'), as well as imperative.

    In Tuyuca, there is no formally or notionally unmarked Declarative, because all the categories not only belong to a single formal system, but also are all notionally evidential. The strongest evidential here is Visual and closest to declarative.

    Again, the declarative may not be the unmarked form. Thus in Huichol, an unmarked form is usually taken as a question, while the declarative has the assertive marker.

    In some languages, there are both a 'weaker' and a 'stronger' declarative. Thus Imbabura has forms for both 'emphatic first-hand information' and 'first hand information'. In Ngiyambaa there are two 'belief' clitics, one representing assertion and the other categorial assertion. Similarly in Hidatsa.



    This chapter deals with Event Modality in different languages. Some of the conclusions reached are as follows.

    Deontic and Dynamic modality refer to events that are not actualized, events that have not taken place, but are merely potential, and may therefore be described as 'event modality'. With dynamic modality the conditioning factors are external to the subject (that is he is permitted, ordered etc to act), whereas with deontic modality they are internal (that is he is able, willing etc to act).

    Commissive (where the speaker guarantees that the action will take place) may also be included under deontic modality.


    The most common types of Deontic modality are the 'directives' where we try to get others to do things. E.g.,

    In English 'may'/'can' and 'must' express two kinds of directives (which also express epistemic Speculative and Deductive).

    You may/can go now. ----Permissive

    You must go now. -----Obligative

    Similar pairs of verbs or constructions can be found in other European languages, e.g., German, Italian, Modern Greek French, Danish, North Frisian etc. In Tamil, the suffixes -laam and -um are used as with epistemic modality.

    Commissive are signalled in English by the modal verb 'shall'. Here the speaker commits himself to ensuring that the event takes place.


    Past tense forms of some of the Deontic modals are used to weaken the force of the modality. 'Must' has 'ought to' and 'should', and 'may' and 'can' have 'might' and 'could' respectively. Notionally, 'should' functions as the modified form of 'must'. 'ought to' and 'should' are essentially conditional referring to what would occur or would have occurred.

    Degrees of obligation are signalled in different ways in different languages. Thus in Albanian, the 'must/should' contrast is marked by different modal verbs.


    There are two types of Dynamic modality, expressing ability and willingness (Abilitive and Volitive), which are expressed in English by 'can' and 'will'. The modal verbs here can also be used for other types of modality.

    In Lisu (Lolo-Burmese) there are distinct forms to indicate two types of ability, one in the sense of 'knowing how', the other in the sense of physical ability. There are three other types of dynamic possibility, indicating freedom from taboo, no hindrance and having sufficient courage.

    Past tense and time

    Past tense forms as with other modals can be used as 'modifications' to express ability and willingness more tentatively, especially when making an offer.

    Imperative and jussive

    Imperative and Jussive may belong to a modal system. Thus in Afar (Cushitic----Ethiopia), there is in the same formal system 'imperative', 'jussive', 'subjunctive', and 'consultative'. All of these could be regarded as 'deontic'--as examples of event modality, relating to possible events in the future. In Ngiyambaa imperative is a member of a mixed system. There are two kinds of imperative in a number of North American languages. In Maidu (North California) Imperative I is used when the action of the order is to be carried out in the presence of the speaker, or when there is no interest in the place of the ordered action.

    Imperative II is used when the ordered action is to be carried out in the absence of the speaker.

    'Purposive' in Australian languages:

    'Purposive' in Australian languages (e.g., in Ngiyambaa) expresses obligation (and epistemic necessity) in main clauses. It can also be used to suggest a result from an unknown cause. In subordinate clauses the same marker is used for both purpose and result. It can also be used to express 'natural result' or 'indirect command'. That is 'purposive' has many functions like the subjunctive in Latin, which can also be used for purpose, result and indirect commands.



    This chapter deals with the similarities in the modal systems, the modal verbs, and other categories.


    The modal systems share a number of features not only in the system themselves, but also for many languages, in the use of modal verbs and the association with possibility and necessity. Notionally, epistemic modality and deontic/dynamic modality seem to have a little in common, yet, in English and many other languages the same form is used for both types. e.g., the following can be interpreted epistemically or deontically. E.g.,

    He may come tomorrow.

    The book should be on the shelf.

    He must be in his office.

    When the same verbs are used for the different types of modality, there are often slight differences in the forms. But that doesn't explain why the forms are basically the same. Even in a Dravidian language Tamil, the two suffixes that are identified as 'permissive' and 'debitive', are used for both epistemic and deontic modality. Similarly in Ngiyambaaa, a single form may be used either deontically or epistemically. Similar examples can be found in Tutatulabal, Cairene Arabic, Abkhaz, Lao, Italian, Lisu etc. The explanation perhaps lies in the notions of 'possibility' and 'necessity', which according to Lyons (1977: 787), are the central notions of traditional modal logic. The terms 'epistemic', 'deontic' and 'dynamic' are taken from a pioneering work on modal logic by Von Wright (1951: 1-2, 28).

    Thus epistemic Speculative and Deductive can be interpreted in terms of what is epistemically possible and what is epistemically necessary.

    John may be in his office.

    =It is epistemically possible that John is in his office.

    John must be in his office.

    =It is epistemically necessary that John is in his office.

    Similarly, deontic Permissive and Obligative can be interpreted in terms of what is deontically possible and deontically necessary.

    You may/can go now

    =It is deontically possible for you to go now.

    You must go now

    =It is deontically necessary for you to go now.

    The importance of possibility and necessity in the modal system is also shown by the distribution of verbs in terms of negation.

    Possibility, Necessity and Negation in English:

    There are two different ways in which a modal expression may be negated, as can be seen in the modal forms of epistemic possibility in English.

    i) Mary may be at school.

    ii) Mary may not be at school.

    iii) Mary can't be at school.

    The difference between the two types of negation can be explained with the notions of possibility and necessity. ii) is to be interpreted as 'possible not', and iii) as 'not possible'. In ii), it is the proposition which is negated, and in iii), it is the negation of modality which is taking place.


    Modal verbs are used in all four of the main types of modality that are found in the modal systems-judgements, evidentials, deontic, and dynamic (though to a very limited degree with evidentials).

    Modal verbs in English

    English has a set of modal verbs that can be formally defined. e.g., 'may', 'can', 'must', 'ought to', 'will' and 'shall' and marginally 'need' and 'dare' (including might, could, would and should).

    They are members of a larger set of auxiliary verbs, which occur with negation, inversion, code, and emphatic affirmation. These are the properties the modal verbs share with the auxiliary verbs 'be' and 'have'. They have the following additional features. E.g.,

    i) Modal verbs do not cooccur;

    ii) They do not have -s forms with 3rd person sg.,e.g., *he oughts to come;

    iii) They have no non-finite forms;

    iv) 'Must' has no morphologically past tense form, although the others do. Of those, only 'could' is used to refer to past time;

    v) There are suppletive negative forms;

    vi) There are formal differences between the modal verbs in their epistemic and deontic senses, in terms of negation and tense.

    Modal verbs in other languages

    Modal verbs exist in other languages too, viz., German, French other Scandinavian languages, Romance languages, even in Mandarin Chinese, and Cashibo (Peru). They have their idiosyncratic properties too besides belonging to the grammatical system.



    Most languages can be characterized as having either the modal system or the mood. However, some languages have both, in two possible ways.

    In a few languages that have a mood system with realis and irrealis markers, the categories associated with irrealis may form a modal system. e.g., Hixkaryana, Serrano, in which the categories marked as irrealis clearly belong to an evidential system. But Central Pomo has a system of evidentials which is independent of it's mood system.

    Romance languages have a system of mood, marked by indicative and subjunctive, but also have a set of modal verbs. However, the modal verbs have not been fully grammaticalized, and in French and Italian, subjunctive is losing ground. In English, the appearance of the modal system (of modal verbs) has been accompanied by the disappearance of the Anglo Saxon mood.

    Thus, the two systems do not coexist in general, and if they do, one is in time, replaced by the other.


    'Will' and 'shall' are formally modal verbs, but are used to refer to future time. They do not often indicate pure futurity, but are usually associated with conditional futures. The connection between future and modality can also be shown historically. There are also plenty of examples of future tenses that are historically derived from subjunctives, eg, some Latin forms. Other languages have future tenses that have their origins in a modal type of auxiliary. It will be seen that Future time is signalled by mood, both by the subjunctive and by the irrealis. This too illustrates the 'potentially modal' characteristics of future time reference.


    The most important relationship between modality and negation is the one that involves possibility and necessity. However, there are other points to note too.

    i) Negation, (also interrogatives) is involved in both modal system and mood;

    ii) Some languages have negative modal verbs, e.g., Latin and Welsh.


    INDICATIVE AND SUBJUNCTIVE: This chapter deals with the various uses of subjunctive in various languages. The category of mood, in European languages particularly, is dealt with in terms of the distinction between the indicative and the subjunctive. It is typically the mood used in the subordinate clauses. Subjunctive can be used in main clauses too. There are six different uses of subjunctive in main clauses in Latin: Jussive ('Imperative'), Volitive (Optative), Obligative (Jussive), Obligative (deliberative), Speculative (potential), Presupposed concessive). Similar examples can be found in Italian. Something similar is found in the West African language Fula.

    Subjunctive in Subordinate clauses:

    The distinction between the indicative and the subjunctive is associated with assertion and non-assertion, and one of the reasons for non-assertion is that the speaker has doubts about the veracity of the proposition.

    Subjunctive is used to indicate what is reported. It can be used in main clauses as well as in subordinate clauses, e.g. German. In subordinate clauses subjunctive is used with verbs of reporting, and belief, and mostly in past tense. Examples can be found in Italian, Latin, Spanish etc. The most common association of Negative with subjunctive is in subordinate clauses where the superordinate clause is negated. In the main clause it is difficult to find a subjunctive with Negation. Some examples can be found in Luvale (Bantu). In subordinate clauses subjunctive is regularly used after negated verbs of belief and report, e.g., Romance languages. The subjunctive is also used with verbs of 'doubt' in the matrix clause, though not under negation. This indicates that subjunctive is licensed in the subordinate clause not by negation in the matrix clause, but by doubt. Interrogatives and Negation often function in very similar ways, and together they can be characterized as 'non-assertive'. Thus verbs of belief and saying in interrogative form induce subjunctive in the subordinate in the subordinate clause in Spanish, Italian, German. In Latin, subjunctive is always required for indirect (reported) questions in Latin, though the indicative is used in direct question. In Spanish, subjunctive is used in the subordinate clause if it is presupposed. Similarly in Italian.

    Futurity is marked as irrealis in languages in which mood is described in terms of realis/irrealis. This is much less common with indicative/subjunctive. In temporal clauses there is wide spread use of subjunctive to refer to hypothetical future events. E.g., Classical Greek and Spanish.

    Subjunctive is used in a conditional clause in a number of languages. In Spanish subjunctive is used for indefiniteness. Similarly in Italian. The subjunctive is also used to express the deontic notions of weak obligation (expressed in English by 'should', the modal past equivalent of 'must'). Similarly, purposives are also marked with subjunctive in many languages. Similarly for Wishes and fears, Resultatives, Imperatives and jussives.

    Subjunctive as Subordinator:

    Though usually there is some notionally irrealis feature associated with subjunctives, a subjuncive can sometimes be simply a marker of subordination. e.g., Latin, Fula, Mangarayi (Australia).


    In this chapter, Palmer investigates the role of realis and irrealis markers in varied languages. Following are some of the findings:

    In many languages, particularly in Native American languages, and those of Papua new Guinea, mood is described in terms of the grammatical markers of realis and irrealis.

    Although the distinction is the same as that between indicative and subjunctive, both being markers of typological categories Realis and Irrealis, there are sufficient differences. Yet there is not always a clear distinction between them.

    Joint and non-joint marking

    There are basically two ways in which realis and irrealis markers function. In some languages, their main function is to cooccur with other grammatical categories. In others they mainly occur in isolation, and are themselves the only markers of specific notional categories.

    E.g., in Amele (Papuan) an irrealis marker is required whenever a future marker is present in the sentence. But in Papuan language, the irrealis marker is itself the indication of future.

    According to Palmer then, the joint markers are those that cooccur obligatorily with other grammatical markers. Both features may appear in a single language.

    In Caddo (Oklahoma), pronominal prefixes of the verb may be distinguished as either realis or irrealis and the choice is determined by the sets of grammatical markers that occur before these pronominal prefixes. Irrealis prefixes are used in conjunction with grammatical markers indicating negation, prohibition, obligation, conditional, etc and also in conjunction with simulative, infrequentative and admirative prefixes. The Caddo system is then predominanatly JOINT.

    The NON-JOINT system is found in Manam (Papuan). The system here is binary. That is, every finite verb must be specified by means of a subject/mood prefix for one of the two moods. Realis is used for i) past events ii) present events & iii) habitual events. Irrealis is used for i) future events ii) commands, exhortations and warnings iii) counterfactual events iv) sequences of customary and habitual events.

    Realis/Irrealis marking is also to be found in subordinate clauses, both oblique clauses and complement clauses (in the same way as indicative/subjunctive). e.g., in mojave the irrealis suffix appears on subordinate verbs to mark unreal or hypothetical situations: desires, conditionals, obligations, counterfactuals, and occassional futures.

    In some languages there are also forms that are unmarked for mood. e.g., In Papuan language Amele, mood is marked only on linked constructions with different subjects. Where there are the same subjects the construction is unmarked for mood.

    There can be wider systems too. E.g., the 'future' 'irrealis' in Kiowa has a number of uses that are associated with irrealis in other languages, but clearly it does not belong to a binary system or even a trinary system (of realis, irrealis and unmarked). Thus there are languages with an 'irrealis' form with several functions. e.g., Maricopa, Mao Naga (Tibeto Burman) etc.

    In Complex Systems like Hixkaryana, there appears to be both an evidential modal system and a system of Realis/Irrealis mood. Mood is indicated by a non-past and a non-past 'uncertain' form of the verb, which can be seen as realis and irrealis. In some languages, the irrealis marker is used, without any other grammatical marker, to indicate a question. e.g., Caddo, Hixkaryana, Serrano, etc.

    Interrogative and Negative are often considered together because they appear to function in similar ways, as in English, where they are both subsumed under 'non-asertion'.

    There are similarities between them w. r. t. realis and irrealis marking. e.g., In Caddo, Mesa Grande Diegueno, Alamblak etc., negatives like interrogatives co-occur with irrealis. There are several instances of the use of irrealis with categories or expressions that are notionally to some degree negative. Thus the infrequentative co-occurs with irrealis in Caddo. In Kiowa, irrealis may indicate that an action might have occurred but did not. Similar examples can be found in other languages too, e.g., Nakanai, Mao Naga etc. Reported, Presupposed, Conditionals etc can be marked as Irrealis in certain languages.

    In Bargam (Papua), habitual past is marked irrealis. Though past time reference is marked realis, the habitual past is not, because it does not relate to specific actions of the past, but to a tendency to act. Similar examples may be found in Kashaya, Tolkapaya Yavapai, Manam etc.

    In Central Pomo, Sursurunga, imperatives and jussives occur with irrealis. Similarly in Takelma. In Kiowa, prohibitive is marked irrealis but not imperative.

    A number of languages make a distinction between 'strong' and 'polite' commands (imperative and jussive) by means of realis/irrealis making. e.g., in Alamblak,

    Nakanai, Jamul Diego etc. In European languages the polite commands are expressed by the subjunctive. In Italian and Spanish even a greater politeness is achieved by using the third person form in addition to the subjunctive.

    Obligative and Abilitive may be marked with an irrealis marker in some languages.


    This chapter deals with the similarities and differences between Subjunctive and Irrealis.

    The terms indicative & subjunctive are traditional terms used in the description of classical and modern languages of Europe, though they have been used for other languages too. The terms realis/irrealis were first used in the description of Australian language Maung. These terms have also been used for Native American languages and those of Pacific, particularly, those of Papua New Guinea.

    Mood in the European languages is a morphosyntactic category closely integrated with person, number tense and voice. That is, it is part of the inflectional system.

    Mood in the realis/irrealis system is often marked by single words or individual affixes or clitics, though there are exceptions too.

    One similarity is that both subjunctive and irrealis markers are often redundant, in that the notional irrealis feature is already marked somewhere in the sentence. But the difference is that a subjunctive is redundant only in the subordinate clause, where the subordinating verb clearly marks the notional feature. By contrast, an irrealis marker

    Commonly occurs in main clauses, where it co-occurs with a grammatical marker that is notionally irrealis.

    At first sight, the notional features associated with subjunctive and with irrealis may appear different, but only if their functions in main clauses are considered. For the notions associated with the subjunctive in main clauses do not include question, denial or futurity (except rarely),

    While these are commonly associated with irrealis. However, in subordinate clauses, these notional features are often associated with subjunctive, often redundantly. However, both irrealis and subjunctive are used to indicate what is presupposed.

    The fundamental question is whether the distinction between Realis and Irrealis in terms of mood, is a coherent and a homogenous category that has typological validity. This is disputed by Bybee et al. 1994. There appears to be four main arguments:

    i) The distinction is rarely realized in a language as a simple binary morphological distinction.

    ii) Irrealis and Subjunctive markers are often semantically redundant in that the meaning is carried by some other element in the context.

    iii) There is a great variation in the notional features marked by them, which makes it difficult to circumscribe a focal meaning for them.

    iv) Some of the notional features appear to be wholly inappropriate.

    Palmer discusses these arguments at length and doesn't agree with them fully.



    Past tense has a modal function expressing 'unreality'. 'Unreality' indicates some degree of lack of confidence by the speaker. Instead of using the term 'unreal', 'modal-tense' is used.

    Examples of Unreal Conditions are conditional sentences.

    In English, there is a contrast between

    If John comes, Bill will leave.

    If John came Bill would leave.

    These are examples of a real and an unreal conditional sentence. With the unreal conditional, the speaker indicates some doubt about the likelihood of the event indicated in the protasis (John's coming), whereas with the real conditional the possibility is left open. The real conditional is expressed by present tense verbs and unreal by past tense verbs.

    Examples of this use of tense for unreality can be found in other languages. Thus in Classical Greek the imperfect indicative is used in both clauses for present unreal conditional. For past unreal conditions it uses the aorist, instead of the imperfect. However where there is clear future reference, real conditionals use the subjunctive (or the future) in the protasis and the future in the apodosis, and the unreal conditionals use the optative in both.

    There is a similar use of tense to mark unreality in the Ethiopian Semitic languages Tigre and Tigrinya, spoken in Eritrea. Similarly in Cushitic language Bilin, unreality is marked by past tense.

    Marking by Tense and Mood:

    Often unreal conditions are marked by both mood and past tense. For unreal conditionals in the present, Latin uses the imperfect subjunctive in both clauses. For unreal conditionals with clear future reference, it has present subjunctive in both clauses. For unreal conditionals in the past, Latin uses a more remote past marker, the pluperfect, but still with the subjunctive. German has imperfect subjunctive in the protasis but either the imperfect subjunctive or the 'conditional tense' in the apodosis. However, the 'conditional' tense is in fact, formed by the imperfect subjunctive of the modal WERDEN, which often marks the future. Spanish also uses 'conditional tenses' and has either the imperfect subjunctive or the conditional in the protasis, but the conditional or the conditional subjunctive in the apodosis.


    Wishes also use past tense forms. Thus Classical Greek uses the optative, the imperfect and the aorist for future, present and past wishes respectively. Similar examples can be found in Latin. English like Classical Greek, sometimes uses an expression that contains the conditional conjunction '-if only'. More commonly it uses the lexical verb 'WISH'. As with conditionals past tense forms are used, but for wishes for the future, 'would' or 'could' are normally used.

    Russian Subjunctive:

    There are other ways in which the past tense is associated with the subjunctive. In Russian, for instance, what is called subjunctive, consists in fact of the particle -by plus the past tense. This is used in several constructions where the subjunctive would be used in other languages.

    Naturally, the modal function of past tense can occur only in languages that have a tense system that differentiates past from present or non- past. It does not occur in languages such as Papua new Guinea, that deal with time relations within the category of Realis/Irrealis mood.

    Bhat (1999: 144) sees the use of past tense for unreal conditionals as a characteristic of 'tense prominent' languages, as distinguished from those that are aspect prominent or mood prominent.

    Palmer poses the obvious question: 'why is past tense, or forms associated with past time so widely used to express this type of modality?

    He discusses the explanations given by different authors, viz., Joos 1964, Steele 1975, Bybee 1995 etc. However, the explanations are not sufficient to explain all kinds of data.

    Critical Evaluation:

    The volume establishes beyond doubt the status of 'Modality' as a Typological category.

    However, a typological study cannot be an end in itself. Major questions come to mind, which can be answered only if detailed syntactic analysis of modality is carried out for each language separately, and comparisons made between different languages at that level. Such a work would be more explanatory and more enlightening. For example, one would like to know,

    1) If typologically there is no distinction between indicative/subjunctive system and realis/irrealis system, how do we explain the structural differences between the two systems?

    2) Admitting that the grammatical markers of modality are very varied, one would like to ask, if there is any similarity between them at the structural level, i.e., in the narrow syntax.

    3) Can a declarative be treated as part of the modal system in all languages, in spite of the fact that in most languages they are not overtly specified for modality?

    4) How can we represent the modality of discourse?

    5) How do we explain the variation amongst languages with regard to what is treated as Irrealis?

    Though it is not at all customary, in my opinion, if books on typological study too include syntactic analysis of at least one of the languages referred to in the study, (for example, it could have been English in this case, since a large part of the data comprises of English data), it would give opportunity to others to work on other languages and to do a comparative syntactic study at the same time.


    Bhat, D. N. S. (1999) The prominence of tense, aspect and mood (Studies in Language Companion Series 49).Amsterdam and Philadelphia: Benjamins.

    Bybee, Joan, Revere D. Perkins, and William Pagliuca. (1994) The evolution of grammar: Tense, aspect and modality in the languages of the world. Chicago: University of Chicago Press.

    Bybee, Joan, and Suzanne Fleischmann, eds. (1995) Modality and grammar in discourse (Typological studies in Language 32.) Amsterdam and Philadelphia: John Benjamins.

    Joos, Martin (1964) The English verb: form and meanings. Madison and Milwaukee: The University of Wisconsin Press.

    Lunn, Patricia V. (1995) The evaluative function of the Spanish subjunctive. In Bybee and Fleischmann 1995:419-49.

    Searle, John R. (1983) Intentionality. Cambridge: Cambridge University Press.

    Steele, Susan (1975) Past and irrealis: just what does it all mean? International Journal of American Linguistics 41: 200-17.

    Von Wright, E. H. (1951): An essay in modal logic. Amsterdam: North Holland.

    SHARBANI BANERJI has submitted her PhD thesis at the Centre for Applied Linguistics and Translation Studies, University Of Hyderabad, India. Earlier she did her M. Phil and M. A Linguistics from JawaharLal Nehru University, New Delhi. Her research interests include Syntax, Semantics and their application in Computational Linguistics. Presently she is working on her book 'Bangla Syntax'. She is looking for work on a (Long Distance) Lexicon Project. net">