Skip to navigation – Site map

Categorization principles of modal meaning categories: a critical assessment

Ilse Depraetere


This paper offers a critical analysis of the features used to define categories of modal meaning (expressed by modal verbs). First, Palmer’s (1990) and Huddleston and Pullum et al’s (2002) definition of dynamic modality is compared with the taxonomic approach to that area of modal meaning in Depraetere and Reed (2011) and Depraetere (2014). The next part of the paper is focussed on the ways in which deontic necessity is defined in Huddleston and Pullum et al (2002), Van der Auwera and Plungian (1998), Coates (1983). Finally, the defining features of the category of dynamic implication are analysed. The discussion is meant to make explicit and assess the diverse strategies used to delineate categories of modal meaning. Some principles of methodological good practice are put forward.

Top of page

Author's notes

I am grateful to the referees, the editors, Susan Reed, Raf Salkie and Christopher Williams for useful input and discussion.

Full text

1. Aims

  • 1 Cf. Van Linden (2012: 12-43) for a useful discussion of the merits and shortcomings of the differen (...)

1In this paper, I will address questions of modal taxonomy, more particularly, the categorization strategies and defining criteria (syntactic, semantic or pragmatic) on which the taxonomies are based. There is disagreement on where category boundaries should be drawn as well as on the number of boundaries to be drawn (and hence the number of classes to be distinguished). The assessment of the categorization principles will be based on three types of comparison: first, the approach to dynamic modality in Palmer (1990) and Huddleston and Pullum et al (2002) will be compared and it will be shown in what ways Depraetere and Reed’s (2011) and Depraetere’s (2014) alternative analysis of that realm of modal meaning is more systematic and more explicit. I will then compare the definitions of deontic necessity across three different taxonomies (Coates 1983, Van der Auwera and Plungian 1998, Huddleston and Pullum et al 2002) and I will argue that rather than pinpointing a taxonomic distinction, deontic necessity refers to a (pragmatic) context of use. Thirdly, I will assess whether the relative absence of the category of dynamic implication (Facchinetti 2002, Collins 2009) in taxonomies of modal meaning can be explained in terms of the kind of meaning distinction it captures. Here again, the conclusion will be that dynamic implication is not a taxonomic distinction on a par with categories like permission, ability or epistemic modality. The general aim of the analysis is to shed light on the nature of defining criteria in modal taxonomies. Given this aim, the selection of taxonomies is by no means meant to offer a comprehensive overview of the field1; rather, the taxonomies chosen illustrate different methodological approaches to the classification and delineation of modal meaning categories. From the analysis I draw some general principles of methodological good practice.

2. Dynamic modality

2In this section I will compare Palmer’s (1990) definition of dynamic modality with that of Huddleston and Pullum et al (2002). I will then show that the analysis of the domain of meaning covered by dynamic modality in Depraetere and Reed (2011) and Depraetere (2014) results in more explicit meaning distinctions based on the systematic application of a set of criteria. The description and evaluation of the taxonomies in general necessarily has to be brief in the context of this paper, the main aim being that of comparing categorization strategies and of reflecting on the nature of the criteria that are used to define (the subclasses of) dynamic modality.

2.1 Palmer (1990) and Huddleston and Pullum et al (2002)

3Palmer (1990) distinguishes three major categories of modal meaning:

(a) epistemic modality: essentially makes ‘a judgement about the truth of the proposition’ (1990: 6)

  • 2 At the beginning of the book, Palmer writes that ‘[there] seems to be no necessity modality that co (...)

(b) dynamic modality: ‘concerned with the ability and volition of the subject of the sentence’ (1990: 7)2

(c) deontic modality: ‘concerned with influencing actions, states or events’ (1990: 6)

4The following examples are given to illustrate the different classes of meaning:

(1) You may not like the idea of it, but let me explain. (1990: 51) (epistemic possibility).

(2) You must find it quite strange being back in London. (1990: 53) (epistemic necessity)

(3) I can make or break my life and myself. (1990: 85) (ability)

(4) I know that place. You can get all sorts of things here. (1990: 84) (neutral possibility)

(5) I’m seeing if Methuen will stump up any money to cover the man’s time. (volition)

(6) You know that certain drugs will improve the condition. (1990: 136) (power)

(7) So one kid will say to another, one kid will make a suggestion to another, he’ll say the moon’s further away from the earth than the sun. (1990: 136) (habit)

(8) When this happens, you will see the boat’s speed fall off and you must pay off just a little. (1990: 116) (dynamic necessity)

(9) If you want to recall the doctor, you may do so. (1990: 71) (permission)

(10) You must keep everything to yourself, be discreet. (1990: 32) (deontic necessity)

  • 3 There is no doubt that the overall picture of the modals is extremely ‘messy’ and that the most th (...)

5Palmer’s basic position is that the ‘messy’ nature of modals should not stop one from defining subclasses of meaning3. He is well aware of the challenge involved, witness his extensive comments on examples that are hard to classify. Even though dynamic modality clearly covers the field of non-epistemic non-deontic meaning, Palmer is not entirely clear about the number and exact nature of the subclasses. The following overview of Palmer’s observations on dynamic, circumstantial, neutral and external necessity shows that the author is at pains to pin down the exact relation between these categories:

(11) p.37: ‘neutral’ or ‘circumstantial’ necessity is, in fact, an additional class [table 1], but it will be treated as a subclass of dynamic modality’ [table 2]

Table 1.




Neutral or circumstantial

Table 2.




Neutral or circumstantial

(12) p.32, p.114: ‘compelling circumstances’ = ‘external necessity’

Table 3.




Neutral or circumstantial or external

(13) p.116: ‘a distinction could in fact be drawn between three kinds of [non-epistemic] necessity, deontic (subject oriented), neutral and external’

Table 4.





6Dynamic modality is initially reserved for so-called ‘subject oriented modality’, that is, modality that ‘is concerned with the ability or volition of the subject of the sentence rather than the opinions (epistemic) or attitudes (deontic) of the speaker (and addressee)’ (1990: 36). Subject orientedness is a lot less in the foreground in the discussion of dynamic necessity (cf. footnote 3); the class of dynamic necessity includes examples like (14) and (15), uses which Palmer labels ‘neutral’ and ‘circumstantial’:

(14) I’ve got to be at London airport at fourish. (1990: 114) (‘circumstances compel’)

(15) When this happens you will see the boat’s speed fall off and you must pay off just a little. (1990: 116) (‘general instructions, (…) the speaker is in no way personally involved’) (example (8) repeated here)

7When discussing can, Palmer distinguishes between a ‘subject oriented use’ and a ‘neutral/circumstantial use’. ‘Neutral possibility’ simply indicates ‘that an event is possible’ (1990: 83), the label ‘circumstantial possibility’ being ‘more appropriate if there is a clear indication of the circumstances in which an event is possible’ (1990: 84):

(16) Signs are the only things you can observe. (1990: 83) (neutral)

(17) You can only get the job if you don’t want it. (1990: 84) (circumstantial)

8When it comes to distinguishing ‘mere possibility’ (neutral possibility) from ‘subject oriented possibility’ (ability), Palmer writes that it is not always possible to draw a clear line, and he gives the following examples to illustrate this observation:

(18) One thing you want to avoid, if you possibly can, is a present from my mother (1990: 85)

(19) The people who cannot very easily raise their wages. (1990: 85)

9He points out that there is ‘indeterminacy between neutral dynamic and subject oriented possibility [ability] and also between neutral dynamic and discourse oriented [deontic] necessity’ (1990: 49, cf. also 132). For instance, ‘rules can be seen as either reports of deontic modality or as saying what is dynamically possible or necessary’ (1990: 104). The following is among the examples he cites (cf. also footnote 15):

(20) In the library you can take a book out and keep it out for a whole year unless it is recalled. (1990: 103)

10Whereas the use of may in such a context tilts the scales in favour of a permission interpretation, ‘where can occurs either interpretation [permission and dynamic possibility] is possible’ (1990: 104).

11Palmer’s definition of epistemic modality is straightforward: epistemic modality ‘essentially (…) [makes] a judgement about the truth of the proposition’ (Palmer 1990: 6). As for deontic modality, Palmer uses the notion of subjectivity as defining criterion; deontic modality is ‘usually subjective in that the speaker is the one who obliges, permits, or forbids’ (1990: 9). Epistemic modality is also subjective in the sense that ‘it relates to an inference by the speaker’ (1990: 9).

12For each of the meanings identified, Palmer discusses a number of formal contexts which they are more or less compatible with, such as negative and interrogative sentences. He looks at the co-occurrence of each of the meanings with a passive infinitive and examines how past and future time reference is expressed. The analysis of the empirical data occasionally results in an update of the definitions, and it is at times a challenge to find one’s way in the maze of connected meanings and to pin down the borderlines between them. This short overview allows us to make the following claims about the defining parameters used:

  • Palmer’s taxonomy consists of three distinct classes of modal meaning (as expressed by modals in English)

  • With the exception of subject oriented modality, which is a more formal feature, the defining parameters are basically semantic or pragmatic, subjectivity being crucial.

  • Some of the features are binary (‘subject oriented’ is); others are not: for instance, the spectrum of subjectivity appears to include: ‘subjective’, ‘neutral’ and ‘circumstantial’ (or ‘external’). It is not entirely clear whether ‘neutral’ and ‘circumstantial’ are (different) subclasses of dynamic modality or whether they are both on a par with deontic modality.

  • Specific features appear to be less or more crucial depending on the realm of meaning (necessity or possibility) that is considered. For instance, while subject oriented examples constitute the core of dynamic possibility, subject orientedness is a lot less in the foreground in the case of dynamic necessity.

    • 4 This will also be demonstrated in the discussion of deontic modality in section 3.

    The borderlines of the conceptual fields occasionally shift. For example, the class of dymanic modality is expanded to include examples that are not strictly subject oriented4.

  • Palmer analyzes how the different meanings interact with certain formal contexts (negation, interrogative contexts, influence of voice, temporal markers)

13Huddleston and Pullum et al (2002) essentially adopt Palmer’s three-fold distinction. Dynamic possibility includes the following ‘considerable range of uses’ (2002: 184):

(21) The most we can expect is a slight cut in sales-tax. – ‘What is reasonable or acceptable’ (2002: 184)

(22) Water can still get in. – ‘What is circumstantially possible’ (2002: 184)

(23) These animals can be dangerous. – ‘What is sometimes the case: the ‘existential use’’ (2002: 184)

(24) I can hear something rattling. – ‘Ability’ (2002: 185)

14Like Palmer, Huddleston and Pullum et al write that the clearest cases of dynamic modality are concerned with properties and dispositions of the persons, etc., referred to in the clause, especially by the subject NP.’ (2002: 178), volition (cf. (25)) and ‘propensity’ (cf. (26)) being further illustrations of this basic meaning:

(25) Jill won’t sign the form. (2002: 192)

(26) Oil will float on water. – ‘characteristic or habitual behaviour of animates or general properties of inanimates’ (2002: 194)

15Dynamic necessity is characterized in terms of the same principle. While subject orientedness is the main defining feature, the category equally includes ‘more peripheral’ examples, like (28), ‘involving circumstantial necessity (comparable to the circumstantial possibility of (22))’ (2002: 185) next to prototypical examples like (27):

(27) Ed’s a guy who must always be poking his nose into other people’s business. (2002: 185)

(28) Now that she has lost her job she must live extremely frugally. (2002: 185)

  • 5 Collins works with the same taxonomy but he does not offer a genuine critical evaluation of this ap (...)

16The overview shows that, as in the case of Palmer’s taxonomy, the category of dynamic modality is broader than just subject oriented modality. Dynamic modality basically covers the range of meanings that are clearly not epistemic and not deontic, the defining strategy being that of definition by enumeration5. In the next section, I will present an alternative approach to the analysis of the domain of meaning covered by dynamic modality, the aim of which is to make the meaning distinctions covered by the examples that belong to this domain, more explicit.

2.2 Depraetere and Reed (2011), Depraetere (2014)

17Depraetere and Reed’s discussion of root possibility (2011) starts from the following observation: the rather large group of possibility examples covered by root possibility that is not ability and not permission appears to be too heterogeneous to justify being classed as at best a single category of meaning, at worst, a ‘dustbin category’ for whatever is left when ability and permission are left aside. They reject the category of dynamic possibility and they argue that three criteria (scope, source and potential barrier) are necessary and sufficient to define, in a coherent way, the five subcategories of meaning that together constitute the class of root possibility. They point to the lack of systematic application of defining criteria: for instance, while in the case of ability, the possibility is said to be predicated of the subject referent, who or what the modal meaning is predicated of is a question that is not systematically addressed. A further observation is that relatively few attempts have actually been made to define permission on the basis of clear criteria, the reason no doubt being that, intuitively, it seems clear what it means. The systematic application of the criteria results in a more explicit taxonomy that brings out very clearly what distinguishes one category from another and what certain categories have in common with each other. In other words, unlike in Palmer’s and Huddleston and Pullum et al‘s model, the subcategories of root meaning are defined on the basis of a single set of defining criteria that is applied to all meaning categories. Table 5 is a summary of this approach.

Table 5. Taxonomy of root possibility (Reed and Depraetere 2011)




general situation possibility (GSP)

situation permissibility













potential barrier

- potential


- potential


+ potential barrier

- potential barrier

+ potential


18The following examples illustrate the different categories of meaning:

(29) “Emily and I want to go to Belgium, to learn French. If we can speak French well, then parents will send their children to us to learn that.” (BNC, British National Corpus) (ability)

(30) “No, no, well yes, she would like to meet you, but it’s for the medicine,” said Endill. “She drinks it all the time.” “She must have a strong mouth,” said the Bookman. “I would find it difficult drinking so much pink. Orange is my favourite, when I can get it.” (BNC) (opportunity)

(31) “The maid is to be dismissed without references,” Katherine continued in the same tone of voice. “She is to leave immediately. If her sister so wishes, she to (sic) can leave -- also without references. That will be all -- oh,” she added, as the butler turned to go, “will you have Master Patrick step in please.” (BNC) (permission)

(32) Progress is not inevitable. People forget things. Things can get worse and that's what's happening now. (BNC) (general situation possibility)

(33) We want to do all we can to help you financially during this time and have developed special loan schemes which are available to post graduates. How much can be borrowed will depend on which course is being taken. (BNC) (permissibility)

19A first criterion that is needed is that of scope of the modal meaning. There is narrow scope if we are concerned with whether something is possible for the subject referent to do; the possibility qualifies the subject referent with regard to the situation referred to in the VP (to Y is possible for X to do). There is wide scope if we are concerned with the possibility of an entire situation. (for situation Z to actualise is possible). The following examples illustrate scope distinctions:

(34) Tim can’t hear very well. (narrow scope) (Depraetere and Reed 2008)

(35) This place is so remote that you can’t hear a sound at night. (wide scope) (Depraetere and Reed 2008)

20In (34) the possibility meaning (in this case the ability) is clearly predicated of the subject referent: it is impossible for Tim to do something. The (lack of) possibility is attributed to Tim. A similar paraphrase does not capture the meaning of the sentence in (35). We do not communicate that the subject referent has the inability to hear a sound; rather, we say that a particular situation is impossible, namely ‘one’s hearing a sound’ is impossible.

21Source of the possibility is the second criterion that is used to distinguish root possibility meanings. Source refers to the person or situation that makes it possible for the subject to do something or that makes it possible for a situation to actualize. What is at stake here is whether the source of the possibility is internal to the subject referent or external to the subject referent. In (34), it is clearly a subject-internal (in)capacity that makes it impossible for Tim to hear very well. In the example in (30), the source of the possibility is subject-external, the idea here being that specific circumstances make it possible for the subject referent to do something.

22The third parameter that is necessary to distinguish between the five basic types of root possibility is that of potential barrier. In some cases, the source of the possibility owes its status as source to the fact that it is capable of imposing a barrier to the subject’s performing a particular action or to the actualisation of the situation. When the source of the possibility is capable of imposing a barrier but does not do so, we get a positive permission (or situation permissibility, when the scope is wide) interpretation; when the source is capable of imposing a barrier and does so, we get a ‘lack of permission (or situation permissibility)’ reading. In the definition of permission and situation permissibility, it is not simply that the source is capable of imposing a barrier to actualization; in addition, the source only qualifies as source by virtue of this capacity. In the example in (31), Katherine has source status because she is in a position to block (or not) the actualization of the maid’s sister leaving.

23Depraetere (2014) addresses the question of whether and how the same criteria can be used to distinguish subcategories of non-epistemic necessity. It seems that scope (wide or narrow) and source (subject-internal or subject-external) are needed, while the criterion of potential barrier cannot be adapted and applied to necessity: the very idea of a barrier being potentially imposed is not compatible with the fact that non-epistemic necessity refers to necessary actualization. The sources involved in necessity all impose actualization and their source status can therefore not be due to the fact that they can potentially block actualization. The application of the two criteria results in a three-class taxonomy of root necessity:

Table 6. Taxonomy of root necessity (Depraetere 2014)

  • 6 While the existence of subject oriented necessity meanings (with need and must) has been pointed ou (...)

narrow scope internal necessity6

narrow scope external necessity

general situation necessity (GSN)









24The following examples illustrate the meaning distinctions:

  • 7 ICE-GB refers to the British English component of the International Corpus of English.

(36) Having said that there are one or two things I want to tell you, and for the moment the only way I can express them is through this letter. I would much prefer to tell you one-to-one. Anyway, prepare yourself. You are special, because you are you. You are lovable, and I … You are unique, and talented too. Are you. You probably didn't expect me to say such things Francoise, but I needed to say them. (narrow scope internal necessity) (ICE-GB W1B-008)7

(37) But if you are a responsible company you have to bear in mind the overall health of the company for the good of all staff and hard decisions have to be taken. (narrow scope external necessity)

(38) Two and a half million people are saddled with hire purchase agreements and mortgages they can no longer afford, and although fewer homes are being repossessed the debts eventually have to be paid. (BNC) (general situation necessity)

25By way of a conclusion, the comparison of the taxonomic principles has shown that mainly semantic and pragmatic criteria have been used to define dynamic modality. In most cases, specific bundles of formal features (relating to negation, the use of the passive infinitive, the type of subject, etc.) are associated with specific meanings but they do not in themselves define the meanings. I have argued that the systematic application of a set of clearly defined features is necessary and I have shown that the use of three criteria – scope, source and potential barrier (in the case of possibility) – make the meaning distinctions covered by the concept of ‘dynamic modality’, or, more in general, root or non-epistemic modality, more explicit and more fine-grained:

  • At the level of root possibility, it clearly sets apart ‘subject oriented’ modality (ability, subject-internal) from all the other types of root modality, which are not subject oriented (subject-external). In this way, the conceptual link between the category of permission (which Palmer classifies as deontic modality) and opportunity (which is called dynamic in Palmer’s approach), both of which are subject-external, is brought to the fore. Likewise, general situation possibility and permissibility are subject-external, but the scope of the modality is wide in these examples, unlike in the case of opportunity and ability.

  • The definition of permission is made explicit in terms of the distinguishing feature of ‘potential barrier’. It is one among five classes of root possibility; it does not constitute a distinct taxonomic class on a par with dynamic possibility, namely ‘deontic possibility’ (as is the case in Palmer 1990 and Huddleston and Pullum et al 2002).

  • The taxonomy is more fine-grained too, in that it brings the importance of scope into the picture, making for, on the one hand, the distinction between (a) narrow scope external possibility (opportunity) and wide scope external possibility (general situation possibility), between (b) narrow scope external necessity and wide scope external necessity, and, finally, (c) between narrow scope external permission and wide scope external situation permissibility.

  • There is no longer a need to define by enumeration or to shift borderlines to accommodate examples that are not central members.

  • It is an elegant approach in that the same criteria are applied across possibility and necessity, with ability being the possibility counterpart of narrow scope internal necessity; opportunity corresponding to narrow scope subject-external necessity and general situation possibility to general situation necessity. In a similar way, it reflects the difference between both domains in that there is, on the necessity side, no category corresponding to permission and permissibility. In other words, this approach results in a clear overview of the dissimilarity and the similarity between the different classes of root possibility and root necessity on the one hand, and that between the realm of possibility meaning and necessity meaning on the other.

Table 7. A comparison of the realm of modal possibility with the realm of modal necessity meaning

















26Having discussed dynamic modality, I will critically assess the ways in which another major category of modal meaning, that of deontic necessity, has been defined, observations which will lead me to conclude that ‘deontic necessity’ captures a context of use, rather than a taxonomic distinction.

3. Deontic necessity

27Deontic necessity is a category that features prominently in taxonomies of modal meaning, but both the defining criteria on which it is based (syntactic, semantic and/or pragmatic) and the conceptual field it covers diverge across taxonomies. In some cases ‘deontic necessity’ covers the complete domain of non-epistemic necessity meaning; in other cases, it is a subcategory of non-epistemic necessity (next to dynamic necessity); in still other cases, it is used to indicate a specific context of use. I will first present some of the strategies used to define this category of meaning and I will then evaluate the parameters on which the definitions are based.

3.1 Deontic necessity in three taxonomies

28The category of deontic necessity is indeed far from receiving uniform treatment in the literature. Below I look at the approach taken to this category in three different frameworks, chosen with the aim of highlighting the ways in which the treatment of this subcategory of modal meaning varies: in Huddleston and Pullum et al (2002)’s analysis of English modal verbs, it is one of three major categories of necessity. In Van der Auwera and Plungian (2008)’s (typological) analysis, deontic root necessity is contrasted with non-deontic root necessity. Coates (1983), one of the first extensive corpus studies of modal verbs in English, questions the usefulness of the concept. Each of the approaches will be briefly outlined (3.1.1 to 3.1.3), and in section 3.2 the type of criteria used to define this category will be assessed.

3.1.1 Huddleston and Pullum et al (2002)

29Huddleston and Pullum et al observe that ‘deontic’ is ‘derived from the Greek for “binding”, so that here it is a matter of imposing obligation or prohibition, granting permission and the like. (…) The person, authority, convention, or whatever from whom the obligation, etc., is understood to emanate we refer to as the deontic source’ (2002: 178). On p.182 they equate deontic necessity with strong obligation and observe that deontic necessity is expressed by must. The source behind deontic must is prototypically subjective (the speaker) but it can also be objective (reports of rules and regulations). The modality in (39), for example, is subjective: ‘I impose on myself, or voluntarily accept, an obligation to do something’ (2002: 183-184). When deontic modality is objective, have to, have got to or need are typically used:

(39) I must do something about this leak. (Huddleston and Pullum et al 2002: 183) (subjective)

(40) You must pull your socks up (Huddleston and Pullum et al 2002: 178) (subjective)

(41) We must make an appointment if we want to see the Dean. (Huddleston and Pullum et al 2002: 183) (objective)

30When the necessity is non-epistemic, it is either deontic or dynamic.

(42) Ed’s a guy who must always be poking his nose into other people’s business. (Huddleston and Pullum et al 2002: 185) (example (27) repeated here)

(43) Now that she has lost her job she must live extremely frugally. (Huddleston and Pullum et al 2002: 185) (example (28) repeated here)

31The sentence in (42) represents prototypical dynamic modality because the necessity is ‘a matter of someone’s properties or disposition (…) rather than of being imposed by a deontic source’. Huddleston and Pullum et al point out that (43) is more peripheral to the dynamic category, since it involves circumstantial necessity: ‘It is not obligation imposed by a deontic source, but simply force of circumstances. But must is less likely than have or need in such cases’ (Huddleston and Pullum et al 2002: 185).

32Summing up, in Huddleston and Pullum et al’s account the defining criterion of deontic necessity is imposition by a (typically) subjective or (possibly) objective source. In other words, two parameters are in the foreground:

  • (a) The notion of source is important, even though it is not entirely clear what differentiates the ‘source’ of (objective) deontic necessity in (41) from the ‘force’ of the (objective) dynamic necessity in (43): it seems that in both cases, it is specific circumstances that make it necessary for someone to do something.

    • 8 They observe (2002: 176-77) that strength is a pragmatic notion as well in the sense that there may (...)

    (b) The inclusion of ‘imposition’ in the definition implies that the concept of strength has a role to play. However, the exact nature of strength is not described in detail: a distinction is made (2002: 175) between semantically strong and weak modality, concepts which are, respectively, linked with necessity and possibility. The authors observe that should expresses ‘medium strength’ but they do not explain in detail how differences in strength, within the modal realm of necessity, as expressed by one specific verb (e.g. must), should be understood or measured8.

3.1.2 Van der Auwera and Plungian (1998)

  • 9 Van der Auwera and Plungian address possibility meanings as well as necessity meanings; I will focu (...)

33Van der Auwera and Plungian (1998) consider deontic modality as a subclass of participant-external modality9. Participant-external modality is defined as referring to ‘circumstances that are external to the participant, if any, engaged in the state of affairs and that make this state of affairs either possible or necessary’ (1998: 80). The modality is deontic when the compelling circumstances external to the participant are ‘some person(s), often the speaker, and/or some social or ethical norm(s) permitting or obliging the participant to engage in the state of affairs’ (1998: 81). The following examples are given to illustrate the two categories of necessity (deontic and non-deontic):

(44) To get to the station, you have to take bus 66. (1998: 80) (non-deontic participant-external necessity)

(45) John must leave now. (1998: 81) (deontic participant-external necessity)

  • 10 Van der Auwera and Plungian (1998: 84) observe that the term ‘participant-internal’ has its shortco (...)

34Participant-internal necessity is expressed by need; it expresses ‘necessity internal to a participant engaged in a state of affairs’10:

(46) Boris needs to sleep ten hours every night for him to function properly. (1998: 81) (participant-internal necessity)

3.1.3 Coates (1983)

35Coates (1983) argues that the modal logic term ‘deontic’ (von Wright 1951) is inappropriate for the discussion of necessity meaning in natural languages. Utterances that communicate root (non-epistemic) necessity share a common, basic meaning, ‘it is necessary for’. Any cut-off points within this conceptual domain are bound to be arbitrary. The root meanings of must lie on a cline extending from ‘strong obligation’ (core) to cases at the periphery where the sense of ‘obligation’ is very weak, the paraphrase in that case being ‘it is important that’. Coates observes that it is only when ‘the authority structure is well-defined’ (Coates 1983: 21) (mother speaking to a child, teacher to a pupil) that it is appropriate to characterize the meaning in terms of ‘obligation’ (which she claims others have called ‘deontic’ (Palmer 1979) or ‘discourse-oriented’ (Palmer 1974)):

(47) You must get out of the bath now. (Coates 1983: 21)

  • 11 Coates argues that ‘root have to, unlike root must, is not associated with a continuum of meaning f (...)

36When the participants are less clearly defined (for instance, in formal prose) such an interpretation is less suitable. Strong obligation is ‘subjective’ and ‘performative’ (1983: 33). Cases of strong obligation constitute the native speaker’s psychological prototype (cultural stereotype) of root must, and have the following features (1983: 33, 37)11:

1. Subject is animate

2. Second person subject

3. Main verb is activity verb

4. Speaker involvement

5. Speaker is interested in getting the subject to perform the action

6. Speaker has authority over the subject

7. Paraphrasable by ‘it is obligatory/absolutely essential that’

37Summing up, both Huddleston and Pullum et al (2002) and Van der Auwera and Plungian (1998) consider deontic modality as a subcategory of non-epistemic necessity meaning. But while in Van der Auwera and Plungian’s account non-deontic and deontic necessity together make up participant-external necessity, in Huddleston and Pullum et al’s approach, deontic modality is on a par with dynamic necessity and epistemic necessity.

Table 8. Comparison of conceptual domain covered by deontic necessity in two taxonomies

Huddleston and Pullum et al 2002

dynamic necessity

deontic necessity

Van der Auwera and Plungian 1998

participant-internal necessity (need)

participant-external necessity

non-deontic necessity

deontic necessity

38Coates (1983), finally, rejects the concept as such; at most it can be associated with what she calls the psychologically prototypical use of must. Irrespective of the fact that the conceptual domain covered by the concept is not the same across taxonomies, it is interesting to analyze in more detail the types of criteria that are used to define it.

3.2 Assessment of defining parameters

39In what follows I will critically assess the parameters on which the class of deontic necessity it has based. My conclusion will be similar to that formulated by Coates, namely, that deontic necessity captures a context of use rather than a taxonomic distinction.

3.2.1 Syntactic features

40Even though Coates’s use of the term ‘deontic necessity’ (or rather, ‘strong obligation’) is restricted to a specific type of utterance and does not reflect a major class of modal meaning, one might adopt her approach to the categorization of meanings and try to define deontic modality in terms of semantic and syntactic features. Among the typical features that yield ‘deontic readings’ are the presence of (a) an animate subject and/or (b) a dynamic infinitive in the active form. (cf. 3.1.3)

(48) The present overdraft /…/ must not be increased. ‘Strong, but passive so non-subjective’ (Coates 1983: 39)

  • 12 This diagnostic is in line with an observation made by Kennedy (2002): ‘through the use of the pass (...)

41While the passive voice can be a helpful feature to identify non-deontic meaning12, it does not solve the problem of distinguishing non-deontic vs. deontic meaning in sentences with active infinitives. Moreover, the presence of a passive/active infinitive does not exhaust the formal features that are indicative of (non-) deontic modality. For instance, one might want to consider the effect of a generic subject or that of the perfect infinitive as well:

(49) Applicants must have lived in Britain for more than two years.

42Coates does not claim that formal features suffice to define modal categories, but the fact that she discusses the co-occurrence of the different meanings with syntactic features shows that she is convinced that formal features have a role to play in the definition of the meanings. (cf. Palmer 1990)

3.2.2 Semantic features

a. Paraphrase ‘it is necessary to’ (+ ‘you are obliged to’)

43A semantic feature shared by deontic and non-deontic non-epistemic necessity is that they both express a necessity: ‘it is necessary for X to’ (cf. Coates 1983: 21). In addition, deontic modality seems to have an additional layer of meaning that can be glossed as ‘X is obliged to’. Paraphrases of this kind are suggested in Coates (1983). However, a ‘definition’ along these lines is unlikely to be helpful; this diagnostic test does not always yield straightforward results when applied in empirical analysis:

(50) The Franks did make great efforts to try and govern Brittany, so it must be asked what stood in the way of preventing their rule, what were the limiting factors to Frankish control? (ICE-GB, W1A-003)

44In accordance with the observation made in 3.2.1, the presence of the passive in (50) seems to rule out a deontic interpretation. However, even though there is no authority insisting on the asking, the appropriate paraphrase seems to be ‘so we are obliged to ask...’ or ‘this (circumstance) obliges us to ask...’ rather than ‘so it is necessary to ask what stood in the way of preventing their rule’ (cf. Depraetere and Reed 2006: 288). In other words, the diagnostic test merely shifts the problem: unless the concept of ‘obligation’ can be clearly defined, the use of paraphrases is not an adequate strategy to define (non-)deontic modality.

b. Modal source

  • 13 This can be explained in terms of the general aim of their paper, which is to develop a semantic ma (...)

45Many of the definitions of deontic modality hinge on the type of modal source associated with the necessity. This is the approach taken in Van der Auwera and Plungian (1998) (cf. 3.1.2, ‘some person(s), often the speaker, and/or some social or ethical norm(s)’). While the description in Van der Auwera and Plungian is rather brief and the examples not very numerous13, the dividing lines are clear.

46Type of source is discussed by Huddleston and Pullum et al (2002) as well (cf. 3.1.1, subjective vs. objective, and, more specifically, for deontic modality, a ‘person, authority, convention’) but in the end it does not provide a firm dividing line between deontic and dynamic necessity as both types of modality can originate in an ‘objective’ source/force. My comments on some of Huddleston and Pullum et al’s examples show that their description of the distinguishing source/force may well need to be made more explicit in order to arrive at a clear classification of source types.

  • 14 This is the approach adopted by Van der Auwera and Plungian (1998), who distinguish between partici (...)

47Depraetere and Verhulst (2008) offer a detailed discussion of the concept of source and apply it to a sample of sentences with have to and must. In other words, it is possible to define the different types of modal source explicitly and operationalize this parameter for corpus analysis in order to pin down potential correlations between specific modals and specific sources. However, the key question is whether it is relevant to define the taxonomic class of deontic modality exclusively in terms of the source with which the modality is associated. Do we need the category of ‘deontic’ necessity if the sole purpose is to assign a label to sentences that are associated with a ‘subjective source’, ‘the speaker’, or an ‘ethical norm’ as opposed to sentences that have a ‘circumstantial source’, an ‘objective source’ or a ‘participant-internal source’? Is this criterion sufficient? And from a taxonomic point of view, does one want to generalize the use of the criterion and distinguish, for instance, between sentences that express deontic possibility and non-deontic possibility, or between deontic permission and non-deontic permission, on the basis of source14? Even though one can single out a specific set of examples on the basis of source, the fact that further factors are involved in most definitions of deontic necessity shows that many scholars are convinced that there is more at stake and that source in itself is not sufficient to assign a differentiated semantic status to this subcategory of necessity meaning. Rather, a variety of factors, among which source of the modality, strength of the modality and illocutionary force, seem to be involved. Coates’ observations about performativity may shed further light on the issue.

c. Illocutionary force and strength of the modality

48Coates gives the following corpus example to illustrate ‘strong obligation’, which is subjective and performative:

(51) “You must play this ten times over”, Miss Jarrova would say, pointing with relentless fingers to a jumble of crotchets and quavers. (1993: 34)

  • 15 Palmer (1990) writes that modals that express deontic meaning can be thought of as performative, i. (...)

49Only 1/14 of all root examples in her corpus are of this type. There are indeed relatively few occasions in which the hierarchic relations between discourse participants are such that direct orders are given, and even if the conditions are met, it is usually not considered to be appropriate to order someone to do something in such a blunt way. Deontic modality involves an attempt to influence the behaviour of others. In other words, it is associated with the illocutionary force of a directive, an ‘attempt (…) by the speaker to get the hearer to do something’ (Searle 1979: 13)15. This means that one needs to take into account the discourse situation, that is, the social distance and hierarchic relations between the discourse participants and their impact on the possible effects of non-compliance.

  • 16 In his discussion of the relative strength of modal verbal expressions in legal English, Williams ( (...)

50A related, but not identical feature is that of strength. Depraetere and Verhulst (2008) define strength of the modality partly according to the nature of the source: in the case of objective sources (they use the label ‘discourse-external’), strength depends on the (a) (im)possibility of non-compliance and the (b) gravity of non-compliance; in the case of subjective sources (the speaker in declaratives and the hearer in interrogatives), the hierarchic relation between the discourse participants needs to be taken into account as well. Verhulst et al (2013) show how the criterion of strength can be made operational for corpus analysis16. Strength of the modality and directive force are two distinct features, as will be clear from the following examples:

(52) ‘Do you think it would be sensible for a woman to go at that time of night into the house of an unmarried man?’ she asked. ‘Well, what happened when you got there?’ I asked. ‘I didn’t go,’ she replied. ‘Mrs Lyons!’ ‘I tell you I did not go. Something happened that stopped me from going. I can’t tell you what it was.’ ‘Mrs Lyons,’ I said. ‘If you did not see Sir Charles, you must tell me why. If you do not, it will look very bad for you if I have to go to the police with this new piece of information about the letter.’

(53) If no clinical signs of recovery are apparent, the surgeon must consider intervention before the peripheral mechanoreceptor and effector sites become irreversibly altered. (ICE-GB: W2A-026)

(54) Filial imprinting takes place in many species of birds and mammals. [. . .] Imprinting is adaptive because it enables the young to recognise and follow their parents. They will grow up in a world of many hostile enemies and one or two protective parents. If they are to survive, the young should choose the right animals to follow. (BNC) (Verhulst et al 2013: 2019)

  • 17 Verhulst et al (2013: 220) observe that the assessment of gravity of non-compliance is culture-depe (...)

51In the three preceding examples, the force of the modality is strong: in (52) the modality is strong because non-compliance could result in the addressee being suspected by the police; in (53) the modality is strong because non-compliance will result in critical health conditions17; in (54) there is reference to instinctive behaviour of birds; non-compliance is not an option. However, it is only the utterance in (52) which has the illocutionary force of a directive: the speaker has authority over the addressee(s) and imposes his or her will to get the addressee(s) to undertake some action. (cf. footnote 15) Alternatively, the examples in (55) and (56) have the force of the directive but the modality is not strong:

(55) ‘Did you black-market your Reeboks?’ ’Yes, tovarich.’ Yashkin shook his head. ‘Is that the way for a policeman to behave?’ ‘I can get your son the Reeboks. If you want? They're not too expensive through my contact.’ ‘If it saves standing in queues, why not?’ ‘You must give me the size,’ said Myeloski. ‘Nine. Size nine.’ (BNC W_fict_prose)

(56) Oh were you you must put a telephone number. (ICE-GB: S1A-077)

52In (55), if the addressee does not give the shoe size, the speaker won’t be able to buy Reeboks, an effect which is negative but which will not impact in a major way on the addressee’s life. In (56), the speaker reminds the addressee to put a telephone number in a supporting letter in the course of a conversation at the UCL Careers office; the effect of non-compliance in this case is not life-threatening; its impact is less considerable than in the examples cited in (52) to (54), as it is clear from the context that the number can be added later in case the addressee forgets to do so.

53If deontic modality is to be defined in terms of strong modality, this will also require offering an explanation for examples like the following, in which there is reference to a reported order (57), or in which there is pragmatic weakening (58):

(57) My boss had said you must read those books. (ICE-GB: S1A-016)

(58) You must let me photograph your baby for my magazine. (ICE-GB: S1A-039)

  • 18 Cf. Nuyts, Byloo and Diepeveen (2010) for a very insightful discussion about the link between permi (...)

54The comment on the usefulness and sufficiency of ‘source’ as a defining feature also applies here: is it justified, from a taxonomic perspective, to define a semantic category on the basis of illocutionary force and/or strength of the modality18? It seems that the feature of illocutionary force, in combination with that of strength, will at most define context of use, and pinpoints a subclass of necessity that has one of the discourse participants as a source (the speaker in declaratives and the hearer in interrogatives). This rather narrow conceptual field does not match that associated with deontic necessity in the majority of taxonomies. Even though an approach along these lines is more explicit (it is spelt out in terms of three defining criteria, that is, source, illocutionary force and strength) than Coates’ proposal, it is close to her view, according to which deontic modality is limited to a specific context of use (‘strong obligation’), non-epistemic necessity in itself being an organic whole in her model. I share Coates’s view that it is not felicitous to use the label of ‘deontic necessity’ to identify a meaning type. In section 2.2, I have argued that a more fruitful approach to the categorization of non-epistemic necessity is in terms of source of the modality and scope, which results in a distinction between narrow scope subject-internal necessity, narrow scope subject-external necessity and general situation necessity. While sentences with modals differ in terms of their illocutionary force, the type of speech act in itself does not seem sufficient to posit a separate category of modal meaning. (cf. also section 4).

55By way of a conclusion to this section, a further terminological point should be made. So far, I have been using the terms ‘semantic’ and ‘pragmatic’ without make explicit the type of meaning involved. There is a general understanding that semantic features are context-independent while pragmatic features are context-dependent. Research in pragmatic theory and philosophy of language (Carston 2010, Recanati 2010, Bach 2013) in the last two decades has shed new light on the semantics-pragmatics interface and has shown that the context is also involved in determining the semantics of an utterance. This implies that ‘context-dependency’ is no longer a defining feature of pragmatics only. Every utterance has an illocutionary force; from this point of view, it is a semantic feature even though the specific illocutionary force intended by the speaker has to be recovered in context. These observations explain why ‘illocutionary force’ and ‘strength’ have been discussed under the heading of ‘semantic features’. In the following section, ‘truly pragmatic’ meaning (in the sense of cancellable meaning) will enter the discussion in the form of implicated, indirect speech acts.

4. Dynamic implication

56Having looked at the different ways in which widely used taxonomic classes have been defined, I will now examine in more detail the opposite case, that of a taxonomic class which is less widespread in classifications of modal meaning, namely the category of dynamic implication (Facchinetti 2002, Collins 2009). The meaning distinction that this term captures is real (see discussion below), but ‘dynamic implication’ is not standardly present in taxonomies of modal meaning. Does this imply that researchers have been oblivious of the meaning distinction at stake, or has this area of meaning been dealt with under a different label or, finally, does the term cover shades of meaning that have been wrongly upgraded into a distinct semantic category? As in the previous sections, the main aim of the discussion is to examine the nature of the defining parameters.

57In his discussion of dynamic possibility, Palmer draws attention to examples that do not in the first place communicate possibility or ability but ‘suggest, by implication, that what is possible will, or should, be implemented’ (1990: 86). He gives the following examples:

(59) You can certainly ring me back this afternoon – there might be something. (‘suggests action be taken by the person addressed’)

(60) Do come early and we can have a drink. (‘combines offer and suggestion’)

(61) Yes, we can send you a map, if you wish. (‘to make an offer’)

58Collins (2009) promotes ‘dynamic implication’ into a subtype of dynamic possibility, which he defines as involving ‘a potentiality for action that is the basis for an implied directive speech act’ (2009: 96). He comments that this class requires ‘an extension of the semantic framework into pragmatics’, whereby the literal meaning is preserved, but the meaning intended by the speaker ‘requires reference to its directive illocutionary force’. The following are among the examples with a possibility modal that illustrate the category of ‘dynamic implication’:

  • 19 ICE-AUS refers to the Australian component of the International Corpus of English.

(62) And you may remember that the organisations the republics were in the Soviet Union competed in the recent Winter Olympics under the title Commonwealth of Independent States. (ICE-AUS S2A-027 40)19 (2009: 96) (directive)

(63) Well I can write on your behalf. (ICE-AUS S1A-004 116) (2009: 104) (offer)

(63) In your letter to me you say that “it is not the ownership of the NRMA that is under review.” Could you explain that to me please? (ICE-AUS W1B-026 86) (2009: 116) (request)

(64) Now if that is the form that your publication is is going to take um then one of the things that you might also think about as an adjunct to your address is the use of audio-visuals or at least visuals. (ICE-AUS S1B 003- 220) (2009: 116) (suggestion)

  • 20 Collins classifies (62) in terms of dynamic possibility ; it seems though that this example illustr (...)

59Facchinetti (2002: 237) goes one step further and considers ‘implication’ to be a taxonomic class on a par with ability, dynamic possibility, epistemic and deontic modality. She argues that the ‘semantic label’ of dynamic modality is only superficially correct: while the semantic load of the modals is dynamic, ‘their correct interpretation depends on their illocutionary force’ (2002: 237). Facchinetti and Collins offer a correct description of the nature of the meaning distinctions: the modal communicates non-epistemic possibility20, but the illocutionary force of the utterance in which it features is foregrounded, which has the effect of bleaching the semantic content of the modal. However, it remains to be seen whether the foregrounding of illocutionary force justifies the creation of a – basically - semantic class of meaning (a subtype of dynamic possibility in Collins’s model, a subtype of non-epistemic meaning in Facchinetti’s taxonomy). This feature sets a specific set of examples apart, but it seems doubtful, also given our discussion of deontic modality, whether an additional semantic class can be created on the basis of the criterion of illocutionary force. The phenomenon described by Palmer, Facchinetti and Collins is mentioned in other analyses of modal verbs (e.g. Quirk et al 1985, Westney 1985, Larreya and Rivière 2014) and it has been widely studied in the context of indirect speech acts, Can you pass me the salt? being a frequently cited example (e.g. Groefsema 1992, Stefanowitsch 2003). The speech act discussions are most often aimed at capturing the inferential process involved in the interpretation of the utterance and/or, more generally, at representing and accounting for the (implicated) speech act in a model of meaning. Indirect speech acts are not exclusively associated with sentences with modal verbs. This observation does not in itself exclude the possibility of using illocutionary force as a defining criterion for a semantic class, but this move does not seem founded from a taxonomic point of view. One may indeed wonder whether it is justified to posit a separate semantic class on the basis of indirect illocutionary force, there being no further differentiating elements with other taxonomic classes.

5. Concluding remarks

60The aim of this paper was to get a better insight into the categorization principles that are used when defining modal meaning categories (as expressed by modal verbs). It has become clear that methodologies are quite diverse and this observation obviously begs the question whether there are any lessons to be learnt.

61First, I have argued that, if the aim is that of drawing clear boundaries and of bringing out similarities and dissimilarities among the different classes, it is necessary to apply the same criteria across meaning distinctions. This uniformity of treatment is sometimes lacking. Secondly, the defining criteria that are used should be made as explicit as possible. The analysis of root modality in terms of scope, source and potential barrier has shown how these aims can be achieved. Thirdly, a distinction should be made between features that define a specific semantic category, and features that determine the context of use of specific modals. For instance, in the case of non-epistemic necessity, the variety of modal verbs at the speaker’s disposal (must, have to, need to, be supposed to, have got to, etc.) makes it likely that contextual features (such as strength of the modality, (indirect) illocutionary force, the speaker as the source of the modality or circumstances as the source of the modality) will tip the scales in favour of the use a specific modal, but that in itself does not imply that these features of use necessarily combine to define semantic classes of meaning. While illocutionary force can differentiate utterances, we should be careful when assessing the role it can play in defining modal meanings. It seems that, the rich and sophisticated analysis of modal meanings notwithstanding, further thought should be given to the role of the contextual features just mentioned: do they define categories of modal meaning or rather, do they, given a specific meaning (e.g. general root necessity), drive the choice of specific modal? So while there is a wealth of literature on taxonomies of modal meaning, there is still room for empirical research which takes into account, and explicitly distinguishes, all the meaning components (context-independent, context-dependent semantic, and context-dependent pragmatic) of sentences with modal verbs.

Top of page


Bach, Kent. 2013. “Odds & Ends (loose/deep/dead)”. Available online. (accessed 3 March 2014)

Bybee, Joan and Suzanne Fleischman. (Eds). Modality in Grammar and Discourse. Amsterdam: John Benjamins, 1995.

Carston, Robyn. “Explicit Communication and 'Free' Pragmatic Enrichment”. Explicit Communication: Robyn Carston's Pragmatics. Eds. Belén Soria and Esther Romero. Basingstoke: Palgrave Macmillan, 2010. 21787.

Coates, Jennifer. The Semantics of the Modal Auxiliaries. London and Canberra: Croom Helm, 1983.

Collins, Peter. Modals and Quasi-Modals in English. Amsterdam: Rodopi, 2009.

Depraetere, Ilse. “Modals and Lexically-regulated Saturation”. Journal of Pragmatics 71 (2014): 160177.

Depraetere, Ilse and Bert Cappelle. “Implicated Modal Meaning in Construction Grammar”. (2014) Paper presented at the third ISLE conference, Zurich, 24-27 August.

Depraetere, Ilse and Susan Reed. “Mood and Modality in English”. The Handbook of English Linguistics. Eds. Bas Aarts and April McMahon. Malden, MA: Blackwell Publishing, 2006. 26990.

Depraetere, Ilse and Susan Reed. “Towards a More Explicit Taxonomy of Root Possibility in English”. (2008) Paper presented at ISLE 1, Freiburg, 8-11 October.

Depraetere, Ilse and Susan Reed. “Towards a More Explicit Taxonomy of Root Possibility in English”. English Language and Linguistics 15.1 (2011): 129.

Depraetere, Ilse and An Verhulst. “Source of the Modality: a Reassessment”. English Language and Linguistics 12.1 (2008): 1–25.

Facchinetti, Roberta. “Can and Could in Contemporary British English: a Study of the ICE-GB Corpus”. New Frontiers of Corpus Research. Papers from the Twenty-First International Conference on English Language Research on Computerized Corpora Sydney 2000. Eds. Pam Peters, Peter Collins and Adam Smith. Amsterdam and New York, NY: Rodopi, 2002. 229–46.

Groefsema, Marjolein. “Can You Pass Me the Salt? A Short-circuited Implicature?”, Lingua 87 (1992): 103–135.

Huddleston, Rodney and Geoffrey Pullum et al. The Cambridge Grammar of the English Language. Cambridge: Cambridge University Press, 2002.

Jespersen, Otto. The Philosophy of Grammar. Chicago, IL: University of Chicago Press, [1924] 1992 .

Kennedy, Graeme. “Variation in the Distribution of Modal Verbs in the British National Corpus”. Eds. Randi Reppen, Susan Fitzmaurice and Douglas Biber. Using Corpora to Explore Linguistic Variation. Amsterdam: Benjamins, 2002. 73–90.

Larreya, Paul and Claude Rivière. (4e edition) Grammaire explicative de l’anglais. Paris : Pearson, (2014).

Narrog, Heiko. “Modality, Mood and Change of Modal Meanings. A New Perspective”. Cognitive Linguistics 16.4 (2005): 677–731.

Nordlinger, Rachel and Elizabeth Closs Traugott. “Scope and the Development of Epistemic Modality: Evidence from Ought to”, English Language and Linguistics 1.2 (1997): 295–317.

Nuyts, Jan. “The Modal Confusion. On Terminology and the Concepts behind It”. Eds. Alex Klinge and Hendrik Høeg Müller. Modality: Studies in Form and Function. London: Equinox, 2005. 538.

Nuyts, Jan, Pieter Byloo and Janneke Diepeveen. “On Deontic Modality, Directivity and Mood: the Case of Dutch Moeten en Mogen”. Journal of Pragmatics 41.1 (2010): 16-34.

Palmer, Frank R. The English Verb. London: Longman, 1974.

Palmer, Frank R. Modality and the English Modals. London: Longman, 1979.

Palmer, Frank R. (second edition). Modality and the English Modals. London: Longman, 1990.

Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech and Jan Svartvik. A Comprehensive Grammar of the English Language. London: Longman, 1985.

Recanati, François. Truth-Conditional Pragmatics. Oxford: Oxford University Press, 2010.

Searle, John R. Expression and Meaning. Studies in the Theory of Speech Acts. Cambridge: Cambridge University Press, 1979.

Stefanowitsch, Anatol. “A Construction-based Approach to Indirect Speech Acts”. Eds. Klaus-Uwe Panther and Linda Thornburg. Metonymy and Pragmatic Inferencing. Amsterdam: Benjamins, 2003. 10526.

Van der Auwera, Johan and Vladimir Plungian. “Modality’s Semantic Map”. Linguistic Typology 2 (2008): 79124.

Van Linden, An. Modal Adjectives. English Deontic and Evaluative Constructions in a Synchronic and Diachronic Perspective. Berlin: Mouton de Gruyter, 2012.

Verhulst, An, Ilse Depraetere and Liesbet Heyvaert. “Source and Strength of Modality: An Empirical Study of Root Should, Ought to and Be Supposed to in Present-day British English”. Journal of Pragmatics 55 (2013): 21022.

Von Wright, Georg H. “An Essay in Modal Logic”. Mind 237 (1951): 115.

Westney, Paul. Modals and Periphrastics in English. Tübingen: Niemeyer, 1995.

Williams, Chris. “Fuzziness in English. What Shall We Do with Shall?” Eds. Anne Wagner and Sophie Cacciaguidi. Legal Language and the Search for Clarity. Bern: Peter Lang, 2006. 237263.

Top of page


1 Cf. Van Linden (2012: 12-43) for a useful discussion of the merits and shortcomings of the different subdivisions of the domain of modal meaning as they are presented in modal taxonomies. Narrog (2005) and Nuyts (2005) similarly offer an interesting comparison of the different classifications put forward in the literature.

2 At the beginning of the book, Palmer writes that ‘[there] seems to be no necessity modality that corresponds to ability’ (p.6). He reconsiders this question in chapter 6 and quotes the following example, the only one with must that seems to have this meaning:

[13] Protoplasm, the living substance of all plants, contains nitrogen and the rose trees must absorb this nitrogen in the form of nitrates. (1990: 129-130)

3 There is no doubt that the overall picture of the modals is extremely ‘messy’ and that the most the linguist can do is impose some order, point out some regularities, correspondences, parallelisms. But there is no single simple solution and one can have some sympathy with Ehrman’s (1966: 9) view that we can only arrive at a ‘rather loosely structured set of relationships’’ (Palmer 1990: 49).

4 This will also be demonstrated in the discussion of deontic modality in section 3.

5 Collins works with the same taxonomy but he does not offer a genuine critical evaluation of this approach: ‘The division in both Palmer’s earlier (1990) book and Huddleston and Pullum (2002), is between epistemic, deontic and dynamic modality. The dynamic category includes ability, volition and also non-deontic root (or, circumstantial meanings). The absence of a larger category that subsumes dynamic and deontic modality (i.e. ‘root’, or Palmer’s 2001 ‘event’ modality) leaves open the possibility that there may be relationships between dynamic and epistemic meanings, just as there are between dynamic and deontic meanings. It is this scheme that is adopted in the present study.’ (2009: 23) Unlike Palmer and Huddleston and Pullum et al, Collins offers quantitative overviews of regional differences in the distribution of the meanings (British English, American English, Australian English), of genre differences and occasionally of features (person of subject with deontic must, harmonic expressions and hedges with epistemic must) typically associated with specific categories meaning.

6 While the existence of subject oriented necessity meanings (with need and must) has been pointed out in previous work (e.g. Palmer 1990: 130, Declerck 1991: 386, Quirk et al 1985: 225, Huddleston and Pullum et al 2002: 185), more often that not, it does not result in a separate taxonomic class, matching the mirror category of ‘ability’, which covers the semantic load of narrow scope possibility examples with a subject-internal source. Van der Auwera and Plungian do distinguish a category of ‘participant-internal necessity’ (cf. section 3.1.2)

7 ICE-GB refers to the British English component of the International Corpus of English.

8 They observe (2002: 176-77) that strength is a pragmatic notion as well in the sense that there may be pragmatic weakening (You must have one of these cakes) or pragmatic strengthening (You may leave now). The pragmatic concept of strength is related to indirect speech acts, which are discussed in section 4.

9 Van der Auwera and Plungian address possibility meanings as well as necessity meanings; I will focus on their account of necessity as this is the notion at stake in the discussion in this section.

10 Van der Auwera and Plungian (1998: 84) observe that the term ‘participant-internal’ has its shortcomings, as a proposition may involve more than one ‘participant’ (you and the bus in (44)). ‘Participant-internal’ coincides with ‘subject-internal’ for a language like English.

11 Coates argues that ‘root have to, unlike root must, is not associated with a continuum of meaning from subjective to objective or from strong to weak. The unitary meaning ‘it is necessary for’ applies in all cases. However, this does not mean that the meanings of must and have to do not overlap, since must can express objective meaning. (…) Must, however, occurs rarely in its objective sense’ (Coates 1983: 55). It follows from her observations that have to cannot express core ‘obligation’: ‘with must the speaker has authority, while with have to the authority comes from no particular source’ (Coates 1983: 55).

12 This diagnostic is in line with an observation made by Kennedy (2002): ‘through the use of the passive voice (e.g. Something must be done rather than You must do it) […] external constraints beyond the control of the speaker can also be implied’ (Kennedy 2002: 79).

13 This can be explained in terms of the general aim of their paper, which is to develop a semantic map that represents cross-linguistically valid synchronic and diachronic connections between modal, premodal and postmodal meanings or uses.

14 This is the approach adopted by Van der Auwera and Plungian (1998), who distinguish between participant-internal possibility (which covers dynamic possibility, ability and capacity), participant-external non-deontic possibility and participant-external deontic possibility.

15 Palmer (1990) writes that modals that express deontic meaning can be thought of as performative, i.e. ‘as indications of speech acts’: 'deontic modals signal ‘directives’ (‘where we get them [our hearers] to do things’).’ (1990: 10) (cf. also 1990: 69-70). Palmer adds that ‘the term [directive] is a little inappropriate since giving permission can hardly be described as ‘directing’. (1990: 6) (cf. Searle 1979: 22) In the case of deontic necessity, there is reference to a particular discourse situation in which the speaker is in a position of authority, in which case must can be used ('I’ve been telling Peter, as I’ve been telling several people, you know, ‘You must get into permanent jobs’, and I’ve been urging Peter to go back to school teaching something, where he’s very, very good.’ (1990: 73)). Palmer is led to extend the definition of ‘deontic modality’ to include some uses that are not strictly performative but only ‘relate to performative uses’ (1990: 70). Utterances that refer to rules and regulations, for instance, ‘could be seen as reports of performative acts by people in authority, but they are also often not clearly distinct from dynamic modality’ (1990: 70). He gives the following example to illustrate this use: A spokesman for Devon County Council’s Weights and Measures Department said “Where a landlady says her place is “two minutes from the sea” it must not mean by jet aircraft” (1990: 104). (cf. 2.1)

16 In his discussion of the relative strength of modal verbal expressions in legal English, Williams (2006) argues that shall is the strongest modal of necessity, followed by strong must, medium-strength be to, (non-modal) indicative simple present being the weakest marker of necessity.

17 Verhulst et al (2013: 220) observe that the assessment of gravity of non-compliance is culture-dependent: what is considered to be precious in one society is not necessarily so in another.

18 Cf. Nuyts, Byloo and Diepeveen (2010) for a very insightful discussion about the link between permission/obligation, deontic modality, directive speech acts and performativity.

19 ICE-AUS refers to the Australian component of the International Corpus of English.

20 Collins classifies (62) in terms of dynamic possibility ; it seems though that this example illustrates epistemic meaning.

Top of page


Electronic reference

Ilse Depraetere, « Categorization principles of modal meaning categories: a critical assessment », Anglophonia [Online], 19 | 2015, Online since 15 July 2015, connection on 27 June 2017. URL : ; DOI : 10.4000/anglophonia.476

Top of page

About the author

Ilse Depraetere

Université Nord de France, Lille 3
UMR 8163 STL

Top of page


Licence Creative Commons
Anglophonia – French Journal of English Linguistics est mis à disposition selon les termes de la licence Creative Commons Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International.

Top of page
  • Logo Presses universitaires du Midi