Show Summary Details

Page of

PRINTED FROM OXFORD HANDBOOKS ONLINE (www.oxfordhandbooks.com). © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Handbooks Online for personal use (for details see Privacy Policy and Legal Notice).

date: 22 April 2019

The History of Modality and Mood

Abstract and Keywords

This chapter traces the current understanding of mood (or mode) and modality back in time, in the Western tradition, giving pride of place to F. R. Palmer, G. H. von Wright, I. Kant, Priscian, Quintilian, Dionysius Thrax, Apollonius Dyscolus, and Protagoras. It sketches how there were and still are at least four different notions of mood, the domain of which was progressively taken over by a notion of modality. It also sketches the tradi¬tion of some key elements in the understanding of modality, viz. necessity and possibil¬ity, which were studied from Greek antiquity onward long before they were subsumed under the heading of modality.

Keywords: mode, modus, subjunctive, optative, imperative, irrealis, sentence type, necessity, possi-bility, negation

2.1 Introduction1

Charleston (1941) is a study of the verb in early eighteenth century English and a survey of the grammatical description of that period. Its very last sentence, quoted with approval by Michael (1970: 434), says that “the treatment of moods by these grammarians of the 17th and 18th century shows a confusion and hesitancy which is still to be observed today among modern grammarians”. If one extends the perspective to the entire tradition of the grammarians’ uses of mood concepts, from antiquity to the modern age, and further extends it to include modality concepts (which are historically much rarer), the impression is the same. There have been thousands of grammatical discussions of mood and modality and though some linguists are self-assured and clear, the field as a whole cannot be said to have come to grips with these notions. This avowal has a negative and a positive side. The negative side is obvious: it is sad that after more than 2000 years our discipline has not reached a better understanding of what is fundamental to modality and mood. On the positive side, one gets the feeling that the subject matter of modality and mood is a fascinatingly difficult one and also that one can still learn from past scholarship.

In this chapter we trace the history of mood and modality in some detail. The overview is restricted to Western linguistics, with its roots in Greek and Latin antiquity (cp. also Kürschner 1987; Malter 2004; Załęska 2004).2 We deal with both the semantic and (p. 10) the formal dimensions of these notions, though we focus on the semantics, and we also show how some of the issues treated in older traditions remain relevant today. Section 2.3 deals with mood and section 2.4 with modality. Section 2.5 comments on the present-day use of these notions. But first, we briefly deal with the etymology of the terms “mood” and “modality”. Note that this chapter, unlike most others in this volume, is more about mood than about modality—contrary to the current situation, for most of the historical periods treated here mood is the prominent notion.

2.2 Terminology and etymology

The English word “mood” has an interesting etymology. There is the Germanic sense of “frame of mind, disposition”, as in to be in a good mood. Next to “mood” there is also a Latin- and French-based word “mode”. The “mode” sense can be paraphrased as “manner”. The origin is Latin modus ‘measure, manner’, and it entered English in late Middle English, either directly or via French mode. It is the English “mode” word that most directly relates to grammar, as the relevant grammatical distinction typically concerns “manners” or forms of the verb. Yet “mood” came to be associated with grammar too. Using “mood” for grammatical “mode”, the Oxford English Dictionary (OED) notes, was “perhaps reinforced by association” with the Germanic “mental state” sense, since some central modes of the verb are the indicative, the imperative, and the optative and these have long been related to mental states such knowing, wanting, and wishing (see 2.3.2). That is one reason why “mood” acquired a linguistic sense. Another reason is simply that the Middle English spelling moode can be seen as the ancestor of both modern “mood” and “mode”. Thus the oldest attestation in the OED for the linguistic sense of both “mood” and “mode” is the same: “A verbe [… ] is declined wyth moode and tyme wyth oute case” (OED, lemma mood & lemma mode). Having two terms for one phenomenon, one Germanic and the other Romance, is typical for English.3 In current French there is now just mode, though earlier there was also moeuf, also deriving from modus. Modo, again deriving from Latin modus, is used in Portuguese, Spanish, and Italian. Germanic languages such as German, Norwegian, and Swedish use the Latin word modus, and if there is a second term, it tends to be a word that means “manner” (Dutch, e.g., prefers wijs/wijze, a general “manner” word).

In the Middle Ages scholars used modus in more than one sense, and one of them was the sense that we now associate with “modality”. Since Greek antiquity there was a strong philosophical, logical, and also theological interest in the concepts of necessity and possibility, and at least since the eleventh century characterizing a proposition as either necessary or possible was said to characterize its modus.

(p. 11) The modern term “modality” derives from the postclassical Latin word modalitas. This Latin term was very rare. In the Library of Latin texts—A (a 63 million words corpus of 3,200 Latin texts from antiquity to the twentieth century) the nominative singular modalitas has only six attestations, all in texts of the thirteenth-century Catalan philosopher Raimundus Lullus. The term modality entered English from French modalité. In the earliest citation in the OED (dating 1545) it has the general meaning “those aspects of a thing which relate to its mode”. Its current linguistic use is recent; the earliest attestation is from 1907. But the linguistic sense immediately relates to logical and philosophical uses, which concern the qualification of a proposition as necessary or possible. For these senses the earliest OED entry dates from 1628.

Interestingly, whereas in Germanic and Romance the term “modality” is clearly related to the term “mood”, in some languages the equivalents of these two terms are unrelated, with the “mood” term being the oldest one. Thus Russian and Greek have, respectively, naklonenie and enklisi (lit. ‘inclination’, see 2.3.2) for “mood”, but for “modality” they use modal’nost’ and tropikotita (lit. ‘tropicality’, based on trópos, which was the classical Greek source for one of the Latin modus senses—see 2.3.4).

2.3 Mood and mode

2.3.1 Quintilian’s modus

The first occurrence of a “mood/mode” etymon in the linguistic sense is commonly traced back (e.g. Wackernagel 1926: 210) to the first century ad and the first of the 12 books of the Institutio oratoria ‘Institute of oratory’, written by the Roman rhetorician Quintilian. In the relevant passage Quintilian discussed grammatical mistakes and he noted that speakers are particularly prone to make errors with respect to verbal categories. One of these categories is the modus.

About substitution, that is when one word is used instead of another, there is no dispute. It is an error which we may detect in connexion with all the parts of speech, but most frequently in the verb, because it has greater variety than any other: consequently in connexion with the verb we get solecisms of gender, tense, person and mood [modos] (or “states” [status] or “qualities” [qualitates] if you prefer either of these terms), be these types of error six in number, as some assert, or eight as is insisted by others (for the number of the forms of solecism will depend on the number of subdivisions which you assign to the parts of speech of which we have just spoken).

(Quintilian, website accessed November 14, 2011, Latin terms added from original)

Interestingly, Quintilian did not tell us here—or anywhere else in his Institutio oratoria—what a mood is. He did supply two synonyms for modus, viz. status and qualitas, both of (p. 12) which did not have the same lasting success as modus, and he also made clear that there was a controversy about how many moods there are: some would accept six and others eight. As in grammar in general, Latin grammars were based on Greek grammars. And so modus, status, and qualitas are translations of Greek. This takes us to Alexandria’s second century bc, its grammarian Dionysius Thrax, and to the Greek word for mood, viz. enklísis.

2.3.2 The Greek grammarians Dionysius Thrax, Protagoras, and Apollonius Dyscolus

Dionysius Thrax is credited with having written the first grammar in the modern sense that survives to this day, the Techné grammatiké ‘Art of grammar’. It is a grammar of Greek. Chapter 13 is about the morphology of the verb and it says that

The verb has eight types of attributes: mood [enklísis], state, species, shape, number, person, tense, and conjugation.

There are five moods—defining [oristiké], imperative [prostaktiké], optative [euktiké], subjunctive [upotaktiké], and infinitive [aparénfatos].

(Kemp 1986: 354, Greek terms added from original)

Here too there was no definition of mood, but there is at least a classification, and the number of moods was not the six or eight alluded to by Quintilian, but five.

For some specialists the Dionysian listing is the oldest extant listing of moods in the Western tradition. But others go back further, be it in an indirect way. In the third century bc Diogenes Laertius wrote Bioi kai gnōmai tōn en philosofia eudokimēsantōn ‘Lives and opinions of eminent philosophers’ (see Diogenes Laertius, translation Hicks 1950). In book 9, 54, he wrote about the fifth century bc presocratic philosopher Protagoras that

[h]‌e was the first to mark off the parts of discourse into four, namely wish, question, answer, command. Others divide into seven parts, narration, question, answer, command, rehearsal, wish, summoning: these he called the basic forms of speech.

(Diogenes Laertius, translation Hicks 1950: II, 467).

In a footnote the translator Hicks claims that these four notions correspond roughly to the optative, the indicative and the imperative (with the indicative corresponding to two of the Greek terms), and Allan (2001: 343) equates them with optative-subjunctive, interrogative, indicative, and imperative.

If one compares the classifications by Dionysius Thrax and Protagoras, one notices an overlap (Table 2.1). But the overlap is by no means complete. None of the Greek terms are the same, and in the English translations only the imperative is exactly parallel. (p. 13)

Table 2.1 Protagoras and Dionysius Thrax on mood

Protagoras

Dionysius Thrax

Hicks (1950)

Allan (2001)

Kemp (1986)

erōtēsis

indicative

interrogative

apokrisis

indicative

defining

oristiké

euchōlē

optative

optative-subjunctive

optative

euktiké

subjunctive

upotaktiké

entolē

imperative

imperative

imperative

prostaktiké

infinitive

aparénfatos

Whether or not Protagoras and Dionysius Thrax were discussing the same thing is a matter of debate. The translator of Diogenes Laertius, Hicks, for instance, as well as the philosopher Allan and the linguistic historiographer Schmitter (2000: 358), would give positive answers. But Robins (1997), also a linguistic historiographer, would disagree. He mentioned both classifications and whereas he considered Dionysius Thrax to have contributed to our understanding of mood (Robins 1997: 44–45), when he discussesed Protagoras, he avoided the term “mood”: “Protagoras [… ] sets out the different types of sentence in which a general semantic function was associated with a certain grammatical structure, e.g. wish, question, statement, and command” (Robins 1997: 32). For the French specialist Lallot (1989: 162), Protagoras did not deal with mode verbal either, but with sentence types, a phenomenon also referred to with modalité d’énonciation. The fact that Dionysius used the term enklísis also suggests that it was something other than Protagoras’ “part of discourse”, for enklísis was essentially a term in the domain of morphology. It meant “flection” or “derivation” and referred to a formation that is different from the basic form of a category, as Lallot (1989: 162) made clear, and he further pointed out that some early Greek grammarians considered moods as counterparts to what cases are for nouns. For the philosopher Nuchelmans (1973: 30), too, Protagoras was not dealing with moods but with the fundamental “kinds of speech”, which were to become an issue in a tradition of rhetoric, developed by Aristotle and his Peripatetic followers and by Stoic philosophers and rhetoricians. But then, paradoxically, Quintilian, who was responsible for the term modus, was a rhetorician, too: rhetoricians were supposed to be dealing with kinds of speech, but for mood he did not focus on speech. Probably, as Nuchelmans (1973: 102) and earlier also Koppin (1877: 11) concluded, the classification of mood was “at least partially inspired” by the classification of kinds of speech (see also Schmitter 2000: 358–359). Interestingly, this relation and tension was present in later accounts too (see Kahn 2004: 250–251 for the thirteenth century). And even today: it is no coincidence that (p. 14) when König and Siemund (2007: 280–281) discuss the typology of basic sentence types they draw attention to the ambivalent character of the imperative:

Another common difficulty with the view of a paradigmatic opposition between the three basic sentence types under discussion [declarative, interrogative, imperative] is the fact that the imperative is often expressed by a specific inflectional form even in languages which do not distinguish the two other types by morphological means. In such languages the imperative is often analysed as being one option in a system of “mood”, which also includes the categories “indicative”, “subjunctive”, “conditional”, “optative” and perhaps others [… ].

(König and Siemund 2007: 280–281)

It is also the imperative that was common to both Protagoras’ and Dionysius’ classifications.

Note also that as a description of basic sentence types the Protagorean description is fairly good. It is very close to modern classifications (such as König and Siemund 2007). The Dionysian classification is more problematic, however, the main problem being the inclusion of the infinitive, which is nowadays no longer considered to be a category of mood. It is, of course, true that the infinitive is a form of the verb, in other words, a manner or mode of the verb. Dionysius did not tell us whether this mode was to be associated with any semantics. So perhaps mode/mood functioned as a wastebasket category and to the extent that the infinitive does not concern more specific categories such as person or tense, it was possible to accommodate it within a vague category of mood.

For a first account of what the mood categories have in common meaningwise, studies of the history of linguistics take us three to four centuries further (second century ad), still in Alexandria, to another founding father of Western linguistics, viz. Apollonius Dyscolus and his Peri syntaxeōs ‘About the construction’ (Lallot 1997). In this elaborate treatise Apollonius treated enklísis as diathesis tes psukhes ‘diathesis/disposition of the mind’ or ‘mental disposition’ (book I §51, book III §§55, 59). What exactly this meant is not clear (cp. Mazhuga 2014). For Lallot (1997: vol. 2, 37) enklísis was to be understood as “modality”, but this is not too helpful, for what exactly is meant by modality here? What does seem clear is that the meaning or use of enklísis was related to the mental state of a person, who is, at least typically, the speaker. Lallot (1989: 162, 1997: vol. 2, 186) called this the “psychological” interpretation of the term enklísis and he quoted the Byzantine commentator Choeroboscos (who worked between the sixth and tenth centuries), who made the psychological interpretation explicit: “enklísis, that is the preference of the mind, that is to say that to which the mind inclines, in other terms, that to which it leans” (Lallot 1997: vol. 2, 186–187; translation ours). This interpretation, sometimes also expressed in terms of “acts” (Lallot 1997: vol. 1, 23), takes the mood idea back to Protagoras. But the connection with form was present as well: according to the specialists enklísis still also referred to the “configuration of sound which conventionally expresses such an inclination or attitude” (Nuchelmans 1973: 101; see also Lallot 1997: vol. 2, 186).

(p. 15) Apollonius accepted the five Dionysian moods and also the terminology, and for some of the moods he made clear what the mind is disposed to: with the indicative (oristiké or apofantikē) one indicates, with the optative (eutike) one expresses a wish and with the imperative (oristiké) one orders (Lallot 1997: vol. 1, 23). These are indeed the three moods for which it was relatively easy to characterize the disposition of the mind. The infinitive and the subjunctive were more problematic, both from a semantic and a formal point of view. The infinitive was considered neutral or unmarked both with respect to expressing any particular disposition of the mind and having formal marking. Dionysius therefore treated it as the most general, basic mood, as its Greek name made clear (aparemphaton ‘which expresses nothing more’), much like the nominative would be the basic case (Lallot 1989: 163). The problems with the subjunctive were different and they were again both semantic and formal. As to the semantics, according to Apollonius the subjunctive expressed doubt (Lallot 1997: vol. 1: 247), but this is not always the case, for Greek uses the subjunctive for a first person plural exhortative. As to the form, different from the other moods, the subjunctive was claimed to need the subordinating conjunction ean ‘if’, a property which the very term “subjunctive” or, clearer perhaps, “subordinative” refers to. This is strange for two reasons. First, mood was no longer only defined in terms of the form of the verb only. Second, the exhortative use just referred to does not involve the conjunction ean, or any other conjunction, for that matter. Interestingly, Dionysius did provide an analysis of the first person exhortative uses of the subjunctive. It was called a “suggestive” use, and the latter was treated as a suppletive form of the imperative (Lallot 1990, 1997: vol. 2, 220) and we thus see Dionysius dealing with an issue that is still controversial today (cp. van der Auwera et al. 2003). The very issue of whether mood should be defined formally only in terms of verbal morphology is a modern one too. In a recent survey of mood in the languages of Europe (Rothstein and Thieroff eds. 2010) the introductory chapter defines mood as a morphological category of the verb (Thieroff 2010: 3), but the author admits that “certain particles [… ] may contribute to a (morphological) mood category” (Thieroff 2010: 3) and he cannot prevent the authors of the Danish chapter (Christensen and Heltoft 2010: 98) from claiming that in this language word order distinctions have replaced verbal morphology and “are best analyzed as mood systems themselves” (see also Van Olmen 2012).

2.3.3 Greek categories applied to Latin: Priscian

So much for the earliest work on mood, all of which concerned Greek. In the history of Western linguistics this Greek tradition was applied to Latin. This created a specific problem for the analysis of mood, viz. one concerning the optative. We will illustrate this with one of the most influential Roman grammarians, Priscian, who worked in Constantinople in the sixth century and whose most important work is Institutiones grammaticae ‘Grammatical foundations’ (see Baratin et al. eds. 2009). For Priscian, the semantic view of moods was that of Appolonius: moods were the diversae inclinationes animi, varios eios affectus demonstrantes ‘the different inclinations of the mind, (p. 16) demonstrating its various affections’ (Calboli 2009: 317), a definition referring to both meaning and form. And like Appolonius and Dionysius before him, he accepted five moods, viz, the indicative (indicativus, also called definitivus), the imperative (imperativus), the optative (optativus), the subjunctive (subjunctivus), and the infinitive (infinitivus). The problem of the infinitive and the subjunctive was the same as for Appolonius, but now there was also the problem of the optative. While in Greek the optative is a distinctive verbal paradigm, in Latin it is at best just a use of the subjunctive. One could reproach Priscian for having pushed Latin into the mould of Greek (Robins 1997: 73) and for having laid the foundations for a “pedagogical disaster” (Hudson 2014: 240), yet if pride of place is to go to meaning, then attributing a separate mood to the expression of wish is defendable. Furthermore, as mentioned before, the psychological perspective easily takes us back to the Protagorean kinds of speech, and why shouldn’t the expression of wish be allowed as a “kind of speech”? But then, why should one stop at five moods?

Indeed, various ancient grammarians did not stop at five, though not always for semantic reasons. Sometimes an impersonal mood was allowed (Michael 1970: 115) and Nuchelmans (1973: 129) mentioned the fourth-century Marius Victorinus, who accepted ten moods: indicative, imperative, promissive, optative, conjunctive, concessive, infinitive, impersonal, gerund, and hortatory. Nuchelmans (1973: 130) further pointed out that the interrogative was occasionally included, and also the participle, and sometimes a distinction was made between what were called the “conjunctive” and the “subjunctive”.

2.3.4 The Middle Ages and the appearance of different notions of “mood”

Throughout the Middle Ages European scholars wrote grammars of Latin and Greek and, from the twelfth century on when the work of Aristotle became better known and universities developed, a lot of attention went to what could be called the “philosophy of grammar”, a general account of how language relates to mind and reality. This theory was termed “Speculative Grammar”, a combination of Greek and Latin grammar and the Catholic interpretation of Aristotle called “Scholasticism”. With respect to modus as used for “imperative mood” and the like, there was little originality: the medieval scholars essentially followed the classical authors, especially Priscian and Donatus. However, interestingly, in Speculative Grammar the term modus appeared with a new sense, best translated with the English word “mode”. The properties of things are their “modes”; there are modes of their existence, of the mind understanding them, and of the language signifying them. The notion of mode was so important that Speculative Grammarians were also known as modistae or “modists”.

In that period, there were three more uses of the term modus, all found in the field of logic. First, in the study of syllogisms, one sense of modus referred to the characterization of the premises and the conclusion as universal or particular and as positive or (p. 17) negative. The example in (1) is such a syllogism, i.e. a valid modus or way of reasoning with, in this case, positive universal quantifiers.

  1. (1) The History of Modality and Mood

Second, in propositional logic, modus also referred to the structure of argumentation, as seen in calling the argument in (2) a modus (ponendo) ponens argument.

  1. (2) The History of Modality and Mood

Third, the term modus more or less referred to what we would now call “modality” and what translates the Greek trópos. The term was used this way in commentaries on Aristotle ([von] Prantl 1855: 654) and it was attested in Boethius (sixth century; [von] Prantl 1855: 695). A proposition was said to have a propositional content (dictum) such as, for example, that it is raining and of this dictum one can say that it is necessary, possible, impossible, or contingent. These characterizations—and sometimes also the ones with the predicates “true” and “false”—were called modi and when there was such a modus, then the whole proposition was modalis (Bocheński 1961: 182; Spruyt 1994).

This multifunctionality of uses of the term modus is not too surprising, for modus basically just means “manner”. But the multifunctionality is also confusing or, if one takes a positive view, thought provoking. In what follows we will see that scholars confused or attempted to integrate—depending on one’s point of view—modus as in the modus of the imperative and modus as in the characterization of a proposition as necessary.

2.3.5 Mood concepts for and in vernaculars

When humanist scholars turned their attention to the vernacular languages, the antique authors remained important, in part because the humanists usually still also worked on Latin and Greek. A good illustration is the Spanish grammarian Elio Antonio de Nebrija (Zamorano Aguilar 2001). He was the author of the first Spanish grammar, the Gramática de la lengua castellana ‘Grammar of the Castilian language’ (1492), but this work was preceded by a grammar of Latin (Introductiones Latinae ‘Latin introductions’, 1481). In both grammars, the discussion of mood was based on Priscian. Thus the general notion of mood was defined as a category of the verb, with reference to demonstrating, ordering, and wishing. And the listing, as also the one for Spanish, had the five most commonly distinguished moods, i.e., indicative, imperative, optative, subjunctive, and (p. 18) infinitive. The old problems of the infinitive and of distinguishing the subjunctive and the optative remained. And sometimes, even when the grammarian only used the classical five moods, new problems appeared.

Thus the first grammarians of German who wrote in German accepted the five moods, including a subjunctive, but the forms listed were actually German indicatives, not Konjunktiv forms (Jellinek 1914: 313). The reason was, as Jellinek pointed out, that the subjunctive was defined as whatever verb form follows the equivalent of the Latin conjunction cum ‘when’. In German they took wann to be the equivalent of cum but wann is followed by an indicative. Sixteenth-century grammarians of French using the five classical moods had a similar problem, this time with the conditionnel. It related to what is not real, a bit like the subjunctive, but the subjunctive-like mode at their disposal was the optative, and that notion did not help much to characterize the conditionnel (Donzé 1967: 113).

As in the classical sources, there were grammarians who distinguished more than five moods. In England, for example, the Latinists Thomas Linacre and John Colet added a non-classical mood, the potential mood, an addition that was to be followed by later grammarians, both of Latin and of English (see Michael 1970: 424; Vorlat 1975: 329; Padley 1976: 49).

An interesting genre is the grammar of Latin or Greek written in the vernacular. Its interest lies in the fact that grammarians often took an implicit or explicit contrastive perspective, trying to provide equivalents in the vernacular. In that context, more particularly in the famous English grammar of Latin known as “Lily’s Grammar”, a notion of “sign” appeared (Gwosdek ed. 2013: 170). A sign of a mood was a formal marking of the mood other than the morphology of the verb. Of course, accepting markers other than verbal morphology was classical practice. For Greek the conjunctive needed the marker ean ‘if’, the Latin conjunctive and optative were often defined with respect to the conjunction cum ‘when’ and utinam ‘if only’, respectively, but we now see the appearance of other such signs. Thus to was considered the sign of the English infinitive, for the optative we had the signs would God, I pray God, and God graunt, and for the newly added potential Lily’s grammar mentioned the signs can, could, might, should, and ought, i.e., verbs or auxiliaries which we would now call “modal” (Blach ed. 1909: 87; Vorlat 1975: 330). We find them also in the Latin grammar De institutione grammatica libri tres ‘About the grammatical foundation three books’ (1572), written by the Portuguese Jesuit Manuel Álvares (see Schäfer 1993). This grammar also accepted a potential mood, probably due to Linacre (Schäfer 1993: 291). Though the Alvares grammar was written in Latin, the author used paraphrases in Portuguese, and for the potential he glossed them with the verbs dever ‘must’ and poder ‘may, can’. We find the modal verbs in the grammars of the vernaculars, too. In the wake of the Lily grammar of Latin, William Turner wrote a grammar of English (A short grammar for the English tongue, 1710; see Vorlat 1975: 336), in which he accepted the potential and said that “[t]‌he Potential Mood signifieth a power, duty or desire, and hath one of those signs, may, can, might, whould, should, could or ought” (Vorlat 1975: 336). In the wake of the Latin grammar of Latin by Álvares, Bento Pereira wrote a Latin grammar of Portuguese (Ars grammatica pro lingua Lusitania addiscenda ‘The art of grammar for the learning of the Lusitanian language’, 1672; see Schäfer 1993) and for the potential he said that it is expressed by the verb poder (Schäfer 1993: 301).

(p. 19) The acceptance of a potential mood and of verbs as signs of this mood is an indication that the Priscian tradition was gradually losing importance. In a survey of 223 grammars of English that were published between 1586 and 1801 and discussed mood, Michael (1970: 434) distinguished between no fewer than 23 different types of accounts and 19 of the grammars denied English to have any moods at all. Thus 1710 saw a grammarian called James Greenwood declaring that “[i]‌n English there are no Moods, because the Verb has no Diversity of Endings” (Michael 1970: 426). Interestingly, the reluctance to honor the small degree of morphological diversity of English with a mood system was connected, by a certain James Pickbourn, to the richness of its auxiliaries, so that the acceptance of no moods comes very close to an acceptance of an abundance of moods: “The English language may be said [… ] to have as many modes as it has auxiliary verbs: for the compound expressions which they help to form, point out those modifications and circumstances of actions which in other languages are conveyed by modes” (Michael 1970: 427).

The mood descriptions of Latin and Greek also allowed unorthodox treatments. Even Latin, with its verbal morphology, was declared to have no moods, as in the Scholae in liberales artes ‘Schools in the liberal arts’ (1559) by the French grammarian Petrus Ramus or Pierre de la Ramée (Vorlat 1975: 330–332; Padley 1976: 89–90). The reason was that the Priscian inclination of the mind was taken to be a matter of the sentence as a whole—a return, it would seem, to the sentential perspective of Protagoras, except that this perspective was not allowed to avail itself of the term modus. Nor are the Latin verbal distinctions traditionally called “imperative”, “conjunctive”, and “indicative”: Ramus forced them into his tense category. A modern view was presented by Gerard Johannes Vossius in his De arte grammatica ‘About the art of grammar’ (1635; Padley 1976): Latin would only have an imperative, a subjunctive, and an indicative. The optative was banned because it could not be defined in terms of verbal morphology, and the infinitive was not really a mood either, or at best only in a “secondary sense”, for it accepted the mood of its “verb accompaniment”—a view Vossius took from the earlier Julius Caesar Scaliger (Padley 1976: 69, 128).

2.3.6 Universal grammar

In 1660 Antoine Arnauld and Claude Lancelot published, so it is assumed, the Grammaire générale et raisonnée, often referred to as the “Port-Royal” grammar, named after schools with the same name (see Arnauld and Lancelot 1969). Arnauld and Lancelot attempted to describe what is universal to human language and how this universality derives from the nature of the human mind. This was a new genre, but it was based on the grammars that the authors knew best, i.e. those of Latin and Greek and of the Western European vernaculars, especially, in the case of the Port Royal grammar, French. Not surprisingly, therefore, the theory of mood was very much in line with the language-specific grammars and the classifications that had come down from antiquity. It so happens that the Port-Royal grammar dismissed the infinitive, but the acceptance of the infinitive as a separate mode had been discussed (p. 20) in language-specific grammars as well (see also Leclerc 2002). An interesting universal grammar, one among many, was the Hermes by James Harris (1751; see Harris 1993). It illustrated a theory of mood that was more in line with the Protagorean approach in that the interrogative was included as a mood (Harris 1993: 143). Harris was also special for associating moods with text types (Haßler and Neis 2009: 1254). The indicative would be the mode of science, for instance, and his “requisitive” (essentially the imperative) would be the mode of legislature.

The interest in universal grammar also led grammarians to devise universal languages. One of these grammarians was bishop John Wilkins, the author of An essay towards a real character and a philosophical language (Wilkins 1668). Like the descriptive Port Royal grammarians before them, these grammarians started from the languages they knew. Thus Wilkins’ system was essentially based on the grammar of Latin (Padley 1976: 208), but interestingly, as a speaker of a language with little verbal morphology and a perspicuous system of modal verbs, his mood theory owed more to English than to Latin. Moods were said to be either primary or secondary.

The Primary Modes are called by Grammarians, Indicative and Imperative. [… ] /The Secondary Modes [… ] make the Sentence to be (as Logicians call it) a Modal Proposition.

This happens when the Matter in discourse namely, the being, doing or suffering of a thing, is considered not simply by it self, but gradually in its causes from which it proceeds either Contingently or Necessarily.

Then a thing seems to be left as Contingent, when the speaker expresses only the Possibility of it, or his Liberty of it.

  1. 1. The Possibility of a thing, depends on the power of its cause, and be expressed when The History of Modality and Mood by the Particle The History of Modality and Mood

  2. 2. The Liberty of a thing, depends upon a freedom from all Obstacles either within or without, and is usually expressed in our Language, when The History of Modality and Mood by the Particle The History of Modality and Mood

    Then a thing seems to be of Necessity, when the speaker expresseth the resolution of his own will, or some other obligation from without.

  3. 3. The Inclination of the will is expressed, if The History of Modality and Mood by the Particles The History of Modality and Mood

  4. 4. The Necessity of a thing, from some external obligation, whether Natural or Moral which we call duty, is expressed, if The History of Modality and Mood by the particle The History of Modality and Mood

(Wilkins 1668: 315, see also Vorlat 1975: 332–333)

(p. 21) The passage is quoted at length for several reasons. First, it illustrates the continuing attractiveness of including the analysis of modal verbs under the heading of “mood”, which goes back at least a century, with Lily’s grammar allowing modal verbs to be “signs” of mood. This was to remain a feature of English grammar writing until well into the twentieth century, with e.g. Sweet’s [A] New English grammar (1891: 420–482) considering the modal auxiliaries as “periphrastic moods” or Zandvoort’s [A] Handbook of English grammar (1950: 102–105) calling auxiliaries “modal” if their meaning comes close to that treated under “verbal mood”. Second, what is also interesting in Wilkins (1668) is that the modal verbs were not merely listed as additional means of expressing mood, but were considered potentially equivalent to the mood of verbal inflection: “As for that other use of the Imperative Mode, when it signifies Permission; this may be sufficiently expressed by the Secondary Mode of Liberty. You may do it” (Wilkins 1668: 316, see also Vorlat 1975: 332). Third, the account of the modal verbs was explicitly linked up with the logicians’ ideas of necessity and possibility, i.e., the logician’s sense of modus, the modern sense of modality. In the twentieth century the linguist Palmer (1986, 2001) did the same thing (see section 2.3.8). Fourth, it offered a fairly modern sketch of the meanings of the English modal verbs. And fifth, it can be noted that Wilkins only used “mode”, not “mood”. The two terms remain in use today, with “mood” now taking the upper hand. But this is recent. Bloomfield (1933) and Hockett (1958), for instance, only used “mode”, not “mood”.

2.3.7 Kant

While Wilkins implicitly had modality under the rubric of mood, the opposite ordering can also be found, in part as a result of the increasing importance of the notion of “modality” within logic and philosophy. This is largely due to Immanuel Kant and his Kritik der reinen Vernuft ‘Critique of pure reason’ (1781; see Kant 1934). Here Kant used the term Modalität for the modus sense that refers to the necessity and possibility of propositions (Pape 1966: 14–15), i.e. to the Boethian sense. The way he did this is considered to be philosophically innovative, even “Copernican” (Poser 1981: 195). What matters for us is that Kant considered modality to be one of the four classes of categories of human judgment, next to quantity, quality, and relation, each comprising three categories.4 In the case of modality they were the problematical, the assertorical, and the apodeictical:

Problematical judgments are those in which the affirmation or negation is accepted as merely possible (ad libitum). In the assertorical, we regard the proposition as real (true): in the apodeictical, we look on it as necessary.

(Kant 1934: 76)

(p. 22) Some grammarians took the three-division of modality to be the basis for understanding Dionysian mood. Thus, starting from the thesis that the laws of language correspond to those of thought, J. G. Hasse appears to have claimed in 1792 (Versuch einer griechischen und lateinischen Grammatologie ‘Attempt at a Greek and Latin grammar’) that the three types of modality are reflected in the three moods: reality is expressed by the indicative, possibility by the conjunctive, and necessity by the imperative (Koppin 1877: 14; Hale 1906: 194). The general idea of explaining mood in terms of modality had various supporters and versions, and was applied to both classical languages and German (see Naumann 1986: 307–311) and probably other languages.

At the very beginning of the twentieth century the modality-based approach to mood was strongly condemned by the classical scholar William Gardner Hale (1906; see also Kluyver 1911). He characterized the entire nineteenth-century scholarship of mood as one of unscientific “metaphysical syntax”, a view quoted with approval by Jespersen (1924: 319). In his own work (esp. Jespersen 1909–49), Jespersen essentially treated mood as a morphological category of the verb, with meanings that are not defined in terms of modality and with no desire to include the modality of what we call “modal auxiliaries”. He did, of course, discuss English verbs such as must and may. First, they were treated in the chapter on tense (vol. 4, 5–16), because some of the modal auxiliaries were old preterit presents, i.e., verbs with preterit morphology but with present meaning. And second, they were also discussed in the sections on verb complementation, more particularly, the ones dealing with the infinitival objects (vol. 5, 170–185). In Jespersen (1924), the theoretical underpinning of the grammar, the approach was similar: there was no place for “modality”, only for “mood”, though interestingly, Jespersen did admit to discussing a second notion of mood, viz., what he (1924: 319–321) called “notional mood”, and he listed no fewer than 20 such notional moods, expressible by “verbal moods and auxiliaries of various languages”. Here we find he must be rich as an illustration of “necessitative” mood and you may go if you like as an example of “permissive” mood. So must and may entered the discussion of mood, but only through a backdoor, for Jespersen (1924: 320, 321) added that he “cannot, however, attach any great importance” to his listing of notional moods, for “[t]‌here are many ‘moods’ if one leaves the safe ground of verbal forms actually found in a language”.

2.3.8 Enter “irrealis”

A special development was the appearance of the term “irrealis”. In current work it stands in opposition to “realis”, but the term “irrealis” probably arose as a member of a set of three terms, with “potentialis” as the third term, with particular relevance to the study of conditional sentences. The term “potentialis” is relatively old: we have seen the sixteenth-century grammarians Linacre, Colet, and Álvares (see section 2.3.5) adding a potential mood to the list of classical moods. The term “irrealis” is a non-classical Latin adjective that is totally absent from the Library of Latin Texts—A. Possibly through Sapir (1992: 186–187; originally published in 1930), the term “irrealis” (p. 23) emancipated itself into a general term—and a noun—for what is not real (Latin realis, medieval Latin, but initially also without the linguistic sense)—or “not factual” or “not veridical”, often also including what is potential. This usage gained prominence in the last quarter of the twentieth century (cp. the comments by Bybee et al. 1994: 236). Symptomatic of its rise is that Palmer (2001) had a chapter on “Realis and irrealis”, while Palmer (1986) did not. The main attraction, it seems, is that it allows one to conceive of mood as having just two categories, viz. realis vs irrealis, instead of e.g. the five from antiquity.

2.3.9 Conclusion

The history of “mood” concepts is confusing. There are at least three important notions. One is the Protagorean notion: it deals with sentence or speech act types. Related, yet different, is the Dionysian notion: it typically, but not necessarily deals with an “inclination of the mind” as it is marked on the verb. A third notion, related but again different, is the Boethian notion, which refers to the characterization of a proposition in terms of, most prominently, necessity and possibility. Part of the confusion is also due to the interference of the term “modality”, to which we now turn.

2.4 Modality

2.4.1 Modality, but without the term “modality”

We have already mentioned that Kant used the dormant term “modality” in an influential way. However, the conceptual issues of what is arguably the core of modality, viz., the distinction between necessity and possibility, occupied scholars since Greek antiquity and were especially prominent in Aristotle (Seel 1982). A central idea in this tradition, of interest for linguistics, is the “square of oppositions”. In essence it goes back to Aristotle too (De interpretatione 6–7, 17b 17–26) but Apuleius of Madaura in the second century before our era might have given its first diagrammatic representation (see Londey and Johanson 1987). The categorization of the square is held to be true for various sets of categories. The most important application is probably quantification, with notions such as “all” and “some”, but we will deal here only with modality. Figure 2.1 shows the classical version of the “modal square”.

The square has a left side, which is positive, and a right one, which is negative. The positive side has an “A” at the top and an “I” at the bottom, referring to the bold face letters in the word affirmo (Latin for ‘I affirm’). The right side has an “E” and an “O”, referring to the bold faced letters in nego (Latin for ‘I deny’). The top values imply the ones at the bottom. Thus if something is necessary, it is also possible. The values that are diagonally opposed to one another are contradictory: they cannot hold true or false together. (p. 24) Thus a proposition cannot be both necessary and not necessary and it is impossible for something to be neither necessary nor not necessary. The relation between A and E is one of contrariety, which means that an A proposition and an E proposition cannot be true together. When something is necessary, it cannot also be the case that it is impossible. The A and E propositions can both be false though: it is perfectly fine for something to be neither necessary nor impossible and, in that case, one would say that it is possible. The relation between I and O, finally, has been called “subcontrariety”. The idea is that I and O cannot be false together, but that they can be true together. This is not self-evident and it has caused problems, apparently as early as for Aristotle (Horn 1990: 454), but the details need not concern us here.

The History of Modality and MoodClick to view larger

Figure 2.1 The Aristotelian square for modality

Issues relating to the Aristotelian square have been discussed by logicians and philosophers until today, but linguists have also joined the discussion, see e.g. Horn (1990). This means that the conceptual analysis that is embodied in the square is deemed applicable to natural language, which is not surprising, for the logicians and philosophers start from natural language, too, some more explicitly than others—see Uckelman (2008: 390–391) for some notes on a thirteenth-century dispute as to whether modus involved adjectives, adverbs, or both. But either way, for a linguist, too, necessity will imply possibility. Nowadays, pride of place goes to verbs as the exponents of modality, and here too, the linguist will readily agree that there is a sense in which must entails may.

Note that the representation with the square focuses on the interaction of modality and negation. Figure 2.1 highlights the effect of external negation, with negation scoping over modality—with not scoping over necessary in O, and im- scoping over possible in E. Figure 2.2 is a version that adds labels highlighting internal negation (necessary not, possible not).

This kind of interaction has engaged linguists too. Palmer (1979: 7–8), for instance, in the first of a few ground breaking studies of English modality, analyzed (3) as the expression of “possible not” and (4) of “not necessary” and found their equivalence to be explained by logic. (p. 25)

The History of Modality and MoodClick to view larger

Figure 2.2 The Aristotelian square for modality, highlighting internal negation

  1. (3) The History of Modality and Mood

  2. (4) The History of Modality and Mood

Or consider the relation between possible and impossible. They are contradictory. Thus one would expect that not impossible is the same as possible. Linguists have worked on this expectation and one might conclude that possible and not impossible are indeed equivalent on a semantic level, but not on a pragmatic level, see Horn (1991).

2.4.2 Kant, von Wright, and Palmer

With Kant the term “modality” gained some frequency, and it replaced the Boethian sense of “mood”, primarily in philosophy and logic, but also in linguistics, especially but not exclusively in the German tradition. Thus, since at least the end of the nineteenth century, Dutch grammarians used a notion of modality (modaliteit) in the modern sense (e.g. Kluyver 1911; Van Wijk 1931: 143–155 [originally published in 1906]), and one of them, Den Hertog (1973: 116 [original from 1892]), explicitly attributed the term to Kant. Interestingly, there was no clear or strong Kantian effect on English grammar writing: as already mentioned, Jespersen (1909–49, 1924) had no need for the term “modality”.

In early twentieth-century English linguistics the term “modality” was not totally absent, however. An interesting use is found in Sapir (1921). He used “modality”, but only in the sense of “mood” (Sapir 1921: 87–88, 108). Another example is the grammar by Zandvoort. It had a section “Mood and modality” (Zandvoort 1950: 102–105). Most of it concerned mood in the traditional sense (indicative, subjunctive,… ), but he used “modal” when a preterit or an auxiliary has a meaning that comes close to one treated under “mood”. The adjective “modal” was explicitly related to “mode”, which was treated as a synonym of “mood”. Lyons (p. 26) (1968: 307–309) was written in the same spirit. “Mood” was his general term, “modality” did occur, but seemingly only for stylistic variation (cf. Hermerén 1978: 10). The Lyons case is interesting. Less than 10 years later, in his other major textbook, Lyons (1977) spent a chapter on both “modality” and “mood”, with 60 pages for modality, illustrated primarily by modal auxiliaries, beating the 50 pages for mood, primarily dealing with speech acts. It was in this period that modality became important in English linguistics. Lyons was not the only linguist to effectuate this change. Major players were Leech (1969 and following), Halliday (1970a and following), Palmer (1979 and following), and Coates (1983). Each of these authors discussed English auxiliaries such as must and may in terms of a notion of modality. The contrast with e.g. Lebrun (1965) could not be stronger: this was a semantic corpus-based study of may and can, in which the author related his results to the relevant literature, and the book did not contain a single occurrence of the term “modality”. An important catalyst of change was Palmer (1979), and, interestingly, here we again witness linguistics turning to logic and philosophy. This time the source of inspiration was not Kant, but von Wright (1951), which Palmer (1979: 2) considered a “pioneering work on modal logic”. Palmer finds von Wright’s work of great interest, not so much for the technical details of modal logic, but for the distinctions made between types of “modality”. Von Wright had four types: alethic, epistemic, deontic, and existential. The details do not matter here, but two of the four notions—“epistemic modality” and “deontic modality”—are now generally accepted as relevant linguistic categories. Interestingly also, von Wright used the term “modality”, but he interchanged it with “mode”, in the Boethian sense of Latin modus. In Palmer, however, only von Wright’s “modality” was used, for he needed “mood” for something else—the Protagoras–Dionysius sense of  “mood”.

At the time when use of the term “modality” was increasing in England, American writers, somewhat independently, started using the term as well, with, e.g., Fillmore (1972: 23) bipartitioning the sentence into a proposition and modality—in an extremely wide sense “including such modalities on the sentence-as-a-whole as negation, tense, mood and aspect”—and with generative grammarians discussing where the locus of the various types of modals in successive formal modals of grammar would be. It so happens that in English the modal auxiliaries are a particularly thorny subject, both formally—they are special verbs—and semantically—they are highly polyfunctional. And since Anglocentric linguistics dominated the world, the term “modality” was heading for global usage. And, of course, in some cases, the use of “modality” coming from Anglocentric linguistics reinforced the use in another vernacular, such as German or Dutch. To some extent, this happened for French too. Brunot (1922) employed a wide sense of modality (modalité), encompassing both mood (mode) and modal auxiliaries (called auxiliaires de modes). But what was more influential is probably the bipartition of the clause, a little like the theory of Fillmore, proposed by Bally (1965: 36–38, 45–46, 216–218), who availed himself of the dictum vs modus terminology, but who also had a use for modalité and mode, each with a meaning different from modus. The Bally approach was influential in France and beyond, in Russia (through the work of Vinogradov; V. Plungian p.c.) and even as far as in Japan (Larm 2006: 66–68). But in current Japan, like everywhere else, it is the English language way of doing grammar and (p. 27) adopting terminology that is exerting its influence. All of these factors conspire to making “modality” a prominent term in today’s grammar and semantics.

2.5 Modality and mood today, and in this volume

What we see in the history of Western linguistics, simplifying somewhat, is the rise of “modality” at the expense of “mood”. Right now, at the beginning of the twenty-first century, the notion of modality is used widely and in many different ways (see Chapter 3). Its prominence is due at least to Palmer, von Wright, and to Kant’s 1781 work, and it essentially goes back to and replaces the Boethian notion of “mood”. “Modality” also assails the Protagorean notion of “mood”, referring to sentence types and/or speech acts, either to replace it or to include it as one dimension of modality. The first line of thought, that of modality replacing Protagorean mood, typical for nineteenth-century Kant, inspired linguistics but it was not taken up. The second line of thought, with modality including Protagorian mood, is found in current research (e.g. Portner 2009: 6, 258–263), but it is not a mainstream idea. However, Protagorean “mood”, as an independent notion, is still very much alive (see e.g. Stenius 1967; Lyons 1968: 308; Lewis 1970; Kürschner 1987), and this is also the perspective taken in this volume.

In addition, this volume also uses the notion of “mood” in a second way, viz. to refer to these aspects of meaning that are treated under labels such as “subjunctive”, “conjunctive”, and also—as a newcomer—“irrealis”. This usage is in line with a good part of current practice and it goes back to the Dionysian concept of “mood”. But there is a lot of divergence and confusion here. The meaning associated with a notion such as “subjunctive” is typically registered in the morphology of the verb. As a result, many linguists take this link to be definitional and they consider “mood” as meaning that has to be expressed through verbal inflection, or as the verbal inflection that expresses this meaning. In the latter case “mood” becomes a formal category which relates to “modality” in the way that “tense” relates to “time” (see e.g. Eisenberg 1986: 98; Palmer 1986: 21–23, 33; 2001: 4; Bybee et al. 1994: 181).

This brief overview by no means covers all current uses of the term “mood” and “modality”. Bhat (1999), for example, uses “mood” for what most contemporary linguists call “modality” and, in his approach, a traditional “mood” category of the imperative is not taken as a category of “mood”, but as a “speech act” category. For Hengeveld (2004) mood is a formal category only (not restricted to morphology) corresponding to the semantic categories of modality and illocution. Given the confusing history of the last 2,000 years, today’s lack of terminological agreement is not surprising and, sadly, the relief of “mood” in some of its uses by “modality” did not bring about the desirable terminological clarity. This implies that no modern user of the terms “mood” and “modality” can take the terms for granted and that one should always explain what one means. (p. 28)

Notes:

(1) Thanks are due to Aikaterini Chatzopoulou (Chicago), Lars Larm (Lund), and Vladimir Plungian (Moscow).

(2) For an overview of other traditions and interactions with the Western traditions, see Larm (2006, 2009), Narrog (2009b), and Masuoka (2009) for Japanese, and Li (2004: 106–109) for Chinese.

(3) An unsuccessful proposal to have a term based on Greek was made by Hare (1970: 21) with tropic.

(4) The most influential application of Kant’s categories to twentieth-century linguistics is the set of Grice’s (1975) four maxims. Grice kept the labels “quantity”, “quality”, and “relation”, but he replaced “modality” with “manner”, thereby also returning to the original sense of modus. In French, however, Grice’s “manner” is sometimes translated as modalité, competing with manière (e.g. Day 2008: 85).