- Email: [email protected]
Modal marking in Akan: the case of anka
Journal of Pragmatics 37 (2005) 997–1013 www.elsevier.com/locate/pragma
Modal marking in Akan: the case of anka Nana Aba Appiah Amfo Department of Language and Communication Studies, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway Received 6 June 2002; received in revised form 2 December 2004; accepted 14 December 2004
Abstract The current article considers a marker, anka, in Akan (Niger-Congo, Kwa branch), which has a variety of uses. Adopting a relevance-theoretic approach, I argue that anka is a marker with a modal meaning. In spite of the variety of uses of this word, I propose a univocal lexical semantics for it, which will be shown to account for its context-dependent function in all the categories. Taking into consideration the interplay between the encoded meaning of anka and pragmatically derived information in actual speech situations, monosemy is seen as a more economical and psychologically plausible way of accounting for the communicative function of this Akan marker. # 2004 Elsevier B.V. All rights reserved. Keywords: Encoded meaning; Higher-level explicature; Implicature; Modality; Relevance; Ambiguity; Vagueness; Indirect requests; Akan
1. Introduction The focus of this paper is the Akan marker anka. This marker has four disparate uses. Using the framework of relevance theory, my aim in this paper is to look at how closely
Abbreviations: CM, Conditional marker; COMP, Complementizer; COMPL, Completive aspect; CONS, Consecutive; DET, Determiner; FP, Focus particle; FUT, Future tense; H, Hearer/Interlocuteor; IM, Interpretive use marker; MP, Motional prefix; NEG, Negation marker; P, Proposition; PART, Particle; PERF, Perfective aspect; POSS, Possessive pronoun; PROG, Progressive aspect; RED, Reduplication; S, Speaker E-mail address: [email protected]. 0378-2166/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.pragma.2004.12.008
998
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
related these different uses are and if there is a possibility of a form of unification, taking into consideration the fact that ‘‘senses are not to be multiplied beyond necessity’’ (Modified Occam’s Razor: Grice, 1989: 47). First I consider the various ways in which anka is used: (a) as a politeness marker in the performance of indirect requests, (b) as a marker in communicating someone’s unfulfilled intentions and desires in what serves as a kind of counterfactual conditional construction, (c) as a marker communicating the existence of a state of affairs which belongs to the past and has ceased to exist, and finally (d) as a marker which indicates that what is described did not actually happen but almost happened. My next concern is with the lexical meaning that anka encodes. My claim is that anka is vague rather than ambiguous, i.e., it has a univocal lexical meaning, which serves as a basis for context-dependent pragmatic enrichment in various directions. In all its uses it indicates that the proposition expressed represents a state of affairs that is not in existence at the time of utterance. The specific meaning of anka in a given utterance is a function of the interaction of its core meaning with other semantically encoded information as well as pragmatically derived information, which the principle of relevance (Sperber and Wilson, 1995: 271) makes readily available to the recipient. Indeed I do acknowledge the fact that anka interacts regularly with other items in specific utterances to bring out its particular pragmatic relevance in these utterances; however, an attempt to give an account of all such items which anka interacts with goes beyond the scope of this paper and will detract attention from the main focus of this paper.1 My claim is that anka is a modal marker, because it encodes a specific epistemic attitude to the proposition expressed by the clause in which it appears. More particularly, it encodes the assumption that the state of affairs described by the propositional content is not factual at the time of the utterance.
2. Cognitive effects and procedural meaning The analysis made in this paper is couched in the cognitively based inferential theory of communication, called Relevance Theory (Sperber and Wilson, 1986, 1995). According to this framework, an utterance is relevant when it combines with contextual assumptions to yield contextual effects. Such effects can be achieved in three ways. First, a new piece of information is relevant when it combines with existing assumptions to derive contextual implications. Second, a new piece of information may strengthen an existing assumption. Finally, a new piece of information may be relevant when it is inconsistent with an existing assumption and as a result one of the pieces of information will have to be dropped in favour of the other. Utterances are interpreted within specific contexts. Context is quite a contentious concept to deal with in the field of language and communication studies. It is defined differently depending on the perspective taken by a scholar. Duranti and Goodwin, in the 1 However, the contributions of some of these accompanying markers, like se, (see Section 3.2) and na, (Section 3.3) are mentioned, where relevant.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
999
introduction to their edited volume, Rethinking Context: language as an interactive phenomenon, recognize the probable impossibility of providing a ‘‘single, precise and technical definition of context’’ (Duranti and Goodwin, 1992: 2). They admit that researchers need not agree on a single definition of context, as different notions of context may be appropriate to different levels of organization. The chapters of this book are significant, in that there is a demonstration of a shift from the notion of context as a frame that surrounds talk (Goffman, 1974) to one in which talk constitutes a main source for the organization of context. Context according to Sperber and Wilson is psychologically defined as ‘‘a subset of the hearer’s assumptions about the world’’2 (Sperber and Wilson, 1995: 15). This apparent simplicity of their notion of context is an attempt not to favour a variable of context over another. For them, all the ethnographic, social, cultural and participant variables which affect context do so by shaping the assumptions that the interlocutor3 uses in the interpretation of an utterance. In a particular interpretation process, the set of assumptions used is the context of the utterance. But how is the context for an utterance chosen by the interlocutor? Certain linguistic indicators in an utterance direct the interlocutor to a set of assumptions that he brings to bear in the interpretation process.4 Linguistic items of this nature are known in relevance-theoretic terms as encoders of procedural meaning. Being in the minority, such linguistic items are contrasted with the majority set of words and expressions, which encode concepts which the interlocutor utilises in constructing the truth-conditional content of the utterance. This relevance-theoretic procedural/conceptual distinction was introduced by Blakemore (1987) with her re-analysis of some of Grice’s conventional implicatures as semantic constraints on implicatures. Further research (Amfo, 2001; Blakemore, 2000; Blass, 1990; Wilson and Sperber, 2004) has revealed that markers of procedural meaning do not only aid the interpretation process by constraining comprehension at the implicit level; they have the potential of influencing comprehension even explicitly.5 The Akan marker anka that I consider in this paper encodes a procedure by which the utterance which contains anka is to be considered as representing a state of affairs which does not exist at the time of utterance. This state of non-existence becomes part of the assumptions constraining the interpretation of the utterance.
2 The ‘‘hearer’s assumption about the world’’ includes all the pieces of information brought into the interpretation process, including what the hearer believes are the assumptions the speaker had in mind in the production of a particular utterance (cf. Sperber and Wilson, 1995: 15–16.). The goal of a pragmatic theory must be to account for utterance comprehension, a ‘mind-reading’ process based on whatever evidence the speaker’s stimulus gives the interlocutor to derive contextual effects from it. 3 Following from the presumption of optimal relevance (Wilson and Sperber, 2004: 612), the speaker in producing an utterance takes into consideration what she believes to be the assumptions held by the interlocutor via the context. 4 For a more detailed discussion on context as a psychological construct, see Sperber and Wilson, 1995 (1986). 5 Grice took comprehension a step further from simply encoding and decoding by distinguishing between ‘what is said’ and ‘what is meant’ (implicatures). However, his distinction also differs quite significantly from the relevance-theoretic explicit/implicit distinction. For more on this, see Carston, 2002.
1000
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
3. Uses of anka Anka is used in a variety of ways. The examples below illustrate four main areas where it is found:6 1. Politeness: (a) Mo-a-w fufuo? Anka me-di o. You-have-pound fufu? Anka I-eat o. ‘Have you pounded fufu? I would have liked to eat.’ (b) Anka me-pe se me-sre w’akondwa yi kakra. Anka I-like COMP I-beg your’chair this little. ‘I would have liked to borrow your chair for a while.’ (Could I borrow your chair please?) c
2. Unfulfilled intentions and desires: (a) Anka -re-ba ha na -te-e se Kofi n-ni h . Anka he-PROG-come here CONJ he-heard-COMPL COMP Kofi neg-is there. ‘He was coming here, and then he heard Kofi was not here.’ (b) Enye se m’a-nya kataa a, anka me-ba a-be-hwe abofra no. It-NEG-be CM I’PERF-get catarrh CM, anka I-come CONS-MP-look child DEF. ‘Had it not been that I have caught a cold, I would have come to see the baby.’ (c) Se anka e-w se wo-tumi k bi. IM anka it-be COMP you-be.able go some. ‘You know, you should have been able to go along.’ c
c
c
c
c
3. A past state of affairs: (a) Na anka -kyere ade w kurom. Then anka s/he-teach thing be downtown. ‘S/he used to teach downtown.’ (b) Kan no anka Onyankop n ben fam.7 Beginning DET anka God near ground. ‘Formerly God used to be close to the earth.’ c
c
c
4. Acts perceived as almost happening: (a) Ns e yi wow bea no CONJ ne ba ara ma Thorns these RED.prick-COMPL woman DET and her child so-much-so that anka ade re-ye a-t w n so. anka thing PROG-be PERF-fall them top. ‘These thorns pricked the woman and her child so much so that they almost fainted.’ c
c
c
7
c
cc
6
Fufu is a type of West African dish prepared by pounding boiled cassava, yam, plantain or cocoyam. This example is taken from Christaller’s (1933) dictionary.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1001
(b) Kwadwo serew no ara ma anka ne mfe re-bu. Kwadwo laugh.PST her so-much-so that anka his ribs PROG-broke ‘Kwadwo laughed at her so much so that his sides almost broke.’8 3.1. Anka as a politeness marker in indirect requests Obeng (1999)9 categorises Akan requests into direct and indirect ones. Indirect requests can further be either conventional or non-conventional. For conventional indirect requests, Weizman (1989), as quoted in Obeng (ibid.), notes that in English, the speaker makes use of some grammatical or semantic device used conventionally for the purpose of requesting; such devices guide the interlocutor in the recovery process. Among the devices are (a) the grammatical structure of questions with modals such as ‘can’ or ‘could’; and (b) the semantic meaning of apologetic expressions such as ‘please’. Such grammatical or semantic devices to signal the indirectness of a request are found in a number of languages, and my argument in this section is that Akan anka should be considered as such a conventional indirect request marking device; anka is a sort of hedge which tones down an otherwise direct request. Anka lexically encodes the procedure that the proposition encoded by the utterance in which it is contained should be considered as unreal. However, the examples given in (1) are only pragmatically appropriate as indirect requests. The request interpretation is not due to the presence of anka; anka is a semantic device used to mitigate the directness of the request. In (1a), if the speaker had simply said medi o (I will eat), it would still be understood as a request but in this case a direct one.10 (1a) thus serves as an indirect request as a result of the presence of anka. In (1b), the semantic content of the verbs used (that is pe (like) and sre (beg)) contributes to the indirectness, with the presence of anka giving the indirectness a further boost. A characteristic of pragmatic markers is that they very often work in tandem with other linguistic items in an utterance to achieve the desired results; the fact that they are highly context sensitive accounts for the apparent difficulty in their semantic categorisation. There are several ways of achieving indirectness and in (1) it is done by the speaker presenting the situation described as nonexistent due to the lexical meaning of anka. Since in Akan, indirection is one of the ways in which politeness finds expression, anka used in such instances can thus be labelled as a politeness marker. The pragmatically inferred meaning of anka in these examples is reminiscent of the way that Sperber and Wilson (1995) and Clark (1991) handle imperatives. According to these authors, imperatives describe a state of affairs, which is both potential and desirable (to someone, frequently the speaker). In (1a), eating fufu is described as desirable to the speaker, but the proposition expressed describes a state of affairs which is only potential, 8 The examples in (4) are taken from a collection of fables in Akuapem Twi ‘Ananse Akuamoa (1961, Bureau of Ghana Languages). The rest of the examples are in Asante Twi. 9 For a detailed discussion on requests in Akan, see Obeng (1999). 10 Obeng (1999) states that direct requests in Akan are in the form of command sentences, that is, imperatives. This category can, however, be expanded to include declaratives which state the speaker’s desire. Note that I define a direct request as a request performed by virtue of the meaning that the sentence encodes, while the term ‘indirect request’ is reserved for cases where understanding of the illocutionary act as a request depends crucially on the inferential processing informed by the encoded logical form of the sentence.
1002
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
not factual. Even though the sentence without anka still describes a non-factual state of affairs, the speaker’s use of anka is significant in that it signals that he cannot take it for granted that the state of affairs described by the proposition will exist, even after the utterance has been uttered. The speaker’s propositional attitude can be inferred to be one of desire; it also comprises the hope that eating fufu is a potential state of affairs which could be brought into reality by the interlocutor acting in such a way that fruition is accomplished. This process seems to be outside the speaker’s own control, and it is therefore up to the interlocutor to decide what to do in this situation. The utterance is understood as a request, and the use of anka makes a considerable contribution to the polite nature of this request. The state of affairs described in (1b) is likewise unreal in the sense of unfulfilled (but at the same time both potential and desirable). Taken literally, these utterances seem to say that my intention is or was for P to come true, where P unfortunately represents a state of affairs which may not come into existence. However, a native interlocutor will not stop drawing inferences at this point. Every native speaker that I presented with a form similar to those in (1), such as anka mepe se me yusu wo fon no (I had wanted to use your phone) or anka merefa wo pen yi (I had wanted to take your pen), either immediately made an offer (such as ‘here you are’, ‘okay’), or declined, giving an explanation. According to a relevance-theoretic interpretation, an assumption is relevant when it achieves contextual effects in a given context (Sperber and Wilson, 1995:112); however, achieving the required contextual effects demands a processing effort. The smaller the processing effort, the more relevant the assumption, and an added function word like anka, in what is already a grammatical sentence, means added processing cost. Why, then, will a speaker of Akan use anka, when its relevance in the utterance requires extra processing effort, as compared to simply making a direct request (without anka)? The reason is that using anka yields certain contextual effects which would not be achieved if the speaker had chosen to use a linguistic form which simply encodes the assumption of being polite, as does the phrase mepa wo kyew (please). This is a set phrase whose conventional meaning does not allow for much variation in its impact on the comprehension process with regard to context-based factors. In contrast, the inclusion of anka in (1) makes manifest a set of assumptions forming part of the context within which the utterance is expected to be processed. These assumptions may include one or more of the following: (i) (ii) (iii)
S realises that the request is not something that H will normally expect from S. S suspects H may not yield easily due to (i) or for some other reason which will have to be deduced contextually. S acknowledges that S and H are in an (extra-linguistic) environment in which H’s right to exert an influence on the practical consequences of the verbal interaction overrides S’s right to do so.
According to Obeng (ibid.), in Akan society requests are usually not considered impositions on the requestees, due to the society’s collective culture and social interdependence. For this reason, certain (direct) requests, which would be deemed impolite and probably inappropriate, even face-threatening in other societies, are generally not perceived as such in an Akan context. By extension, requests in Akan society, as compared to certain Western
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1003
societies, are non-imposing linguistic acts. This holds true to a large extent for rural Akan communities, where a high level of dependence among members is to be expected. In urban and semi-urban Akan communities (such as Koforidua, where the data in (1) was collected), social interdependence is gradually on the decline and as a result, the way the Akan conceptualise requests may vary, depending on certain socio-pragmatic circumstances. But also here, as Obeng is quick to add (ibid.), the interpersonal relationship between the requester and the requestee should not be overlooked. The requester in (1) certainly ignores his interpersonal relationship with the requestee as well as other socio-pragmatic considerations. An able-bodied person who asks for food (a very basic need) from someone who may only qualify as a mere acquaintance certainly threatens the interlocutor’s negative face. And note that the request was not made because the requester saw the requestee eating or being about to eat; it was made simply on the assumption that the requestee had prepared food. Now, let us see how the above assumptions work with respect to example (1a). This utterance was said by a young man (Y) to a 19-year-old girl (E). They had met once or twice in the past, with minimal interaction having occurred between them. I reproduce (1a) below, including E’s reply as (5). Y began by asking E if her parents were home and then continued: Y: Mo-a-w fufuo? Anka me-di o. You-PERF-pound fufu? Anka I-eat o. ‘Have you pounded fufu? I would have liked to eat.’ E: Eneese wo-a-nya, yen na ye-noa aduane ma wo-di. Then COMP you-PERF-get, we FM we-cook food give you-eat. ‘Oh is that what you think? Is it we who cook for you?’ (Is it our responsibility to cook for you?) c
(5)
Given the not-so-close interpersonal relationship between Y and E, a request for food in the above-mentioned circumstances is not something that E would normally expect from Y, and Y is probably aware that, as a result of this, E will not accede to his request, especially if it is made directly. Even a request that is toned down by the use of anka would, in this particular case, not only come unexpected for E, but appear inappropriate as well, something which her reply confirms. In using anka, Y implies that he is aware of all these assumptions, which would normally have prevented him from making such a request directly. However, since he prefers to disregard the socio-cultural assumptions in order to have his desire fulfilled, he makes the awareness of the social inappropriateness of his request manifest, and hopes that in doing so, he will cause E to yield to his request. Presumably, he felt that this might be the only way to make E recognise his communicative intentions and at the same time give himself a chance to achieve his goal. The use of anka as a politeness marker is further illustrated by the following examples from a radio discussion. This discussion took place during a review programme. The newspaper article under review was supposed to have been based on a Serious Fraud Office (SFO) interim report, which apparently was not in favour of some public official. The report in question had not actually been made public officially, but the newspaper which published the article claimed to have access to it anyway. In (6), one of the panelists appeals to the relevant authorities to make such reports public in order to put an end to speculations,
1004
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
while another panelist in (7) reiterates an official’s request that journalists should exercise some restraint when dealing with reports which are not considered as final.11 Anka me-ka se se SFO ye rep t bi a they should make the Anka I-say COMP CM SFO do report some CM. . .. . . report open to the public. ‘I would request that when the SFO prepares a report, they should make it c
(6)
open to the public.’ Anka -be-sre se anka nsentwerefo ye-n-to yen bo Anka he-FUT-beg COMP anka newswriters we-IMP-throw our chest ansa, na SFO na ew finale. . .e. . . mandate or power before, CONJ SFO FM it-has final mandate or power ‘He would ask that journalists exercise restraint because it is the SFO which has the final mandate (in such matters).’ c
c
(7)
c
In both (6) and (7), the speakers use anka as a downtoning marker while making their requests, since they know that unless the interlocutors act as suggested, there is nothing that they can do practically about these situations. In (7), note that anka is used both in the main clause and in the complement clause. The first anka used in the utterance is an indication that the speaker is making an indirect request, that is, for journalists to take their time in matters of that nature. The second anka directs the interlocutor to a particular implicature,12—that journalists are not taking their time. In other words, the proposition that journalists should take their time (which is the message explicated without anka), is not true at the time of utterance. This particular implicature becomes part of the context within which the utterance is interpreted. This particular usage of anka, where the speaker states what will be desirable if the pragmatic conditions are or (in counterfactual situations) were fulfilled, is comparable to a familiar pattern of the usage in English, where a potentially face-threatening act is prefaced with an expression involving the past tense form of modals. (8)
I would have liked to ask you. . .
The speaker will often times go ahead with her (indirect) question or request in spite of her apparently counterfactual preface, having first expressed her realisation that the question or request might be deemed unsuitable under the circumstances. An utterance like (9), produced by a supervisor addressing a student, would normally be relevant beyond what seems to be literally an example of wishful thinking. (9)
It would have been a good thing if you and I could get together to discuss your paper.
11 It is not uncommon to find code mixing and code switching among educated Akans and I produce the examples as they were recorded. 12 An implicature, according to Carston (2002: 377), is a ‘‘communicated assumption which is derived solely via processes of pragmatic inference’’.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1005
3.2. Anka in communicating unfulfilled intentions and desires Another use of anka is found in utterances expressing intentions or desires which cannot be, or have not been, fulfilled (as exemplified in (2), repeated here for convenience). The reason for the unfulfilled intention or desire is often, though not always, given in a conjoined clause (as in (2a)) or a conditional clause (as in (2b)). The anka-clauses in (2), unlike those in (1), are not interpreted as requests or entreaties, due to the general sentence constructions. The forms in (2) are counterfactual conditionals with the protasis introduced by anka. However as in (1), they describe a desirable and a potential state of affairs. (2) (a) Anka -re-ba ha na -te-e se Kofi n-ni h . Anka he-PROG-come here CONJ he-hear-PST COMP Kofi NEG-is there. ‘He was coming here, but then he heard Kofi was not here.’ (b) E-n-ye se m’a-nya kataa a, anka me-ba a-be-hwe abofra no. It-NEG-be CM I’PERF-get catarrh CM anka I-come CONS-MP-look child DEF. ‘Had it not been that I have caught a cold, I would have come to see the baby.’ (c) Se anka e-w se wo-tumi k bi. IM anka it-be COMP you-be.able go some. ‘You know, you should have been able to go along. c
c
c
c
c
In (2a), the clause expressing the intention starts the utterance, before an explanation is given for the failure to carry out the action. The anka-clause describes a desirable and potential state of affairs, or at least so it seemed to the referent of the pronominal (he), until in the conjoined clause, he was informed of Kofi’s absence, which went against what was implicitly assumed in the protasis. The relevance of the anka-construction is that while it expresses a potential and desirable state of affairs, it at the same time implicates that the actual performance of this state of affairs is hindered by something. What anka does is to make more manifest a specific implicature of (2a). As the state of affairs referred to in the clause modified by anka does not exist at the time of utterance, and as the state of affairs described would have belonged to the future in any case, the inferred meaning is that the referent of (he) is not coming. In (2b), the reason for the inability to carry out the intention is given before the intention is expressed. The situation described in the main clause is desirable, and it would also have been potential if it were not for the fact that the negative proposition expressed in the preceding conditional clause is presented as counterfactual (due to the co-occurrence of enye ‘it is not’, and the conditional marker se. . .a). In (2c), the assumption expressed is that the potential and desirable state of affairs described by that utterance is indeed desirable to the speaker, and is furthermore presented by the speaker as being probably desirable to the interlocutor as well. This is in spite of some apparent contrary evidence, which the interactants may have taken to imply that the state of affairs cannot be fulfilled. (2c) was said by a wife to a husband concerning a men’s camp meeting organised by their church. This meeting has not yet taken place, but both the speaker (wife) and the interlocutor (husband) are aware that the interlocutor does not intend c
c
1006
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
to go; this explains the use of the first se in the utterance. This se is not a complementiser (as is the second se in this utterance). It is a marker which indicates that the information that I’m about to share is not new information, and you know that as well as I do. What is explicitly communicated in (2c) is that the interlocutor should be able to go. The combination of se and anka guides the interlocutor to the assumption in (10), which forms part of the context within which (2c) is processed: (10)
S knows that I know that she is aware I am not going, but it is desirable (to her) that I go, and both of us know that.
When anka is used in these instances, it gives rise to the contextual assumptions that some desire or potential cannot be achieved. The following two examples, said of a president of the National Union of Ghana Students (NUGS) who lost his life in a car accident just a month after his inauguration. again illustrates the use of anka in unfulfilled desires and potentials. (11) Anka -re-k Tamale a fofie13 an-to Atta na abrantie Anka he-PROG-go Tamale PART sacred-day NEG-arrive twin CONJ gentleman yi a-k a-k -ye Nyankop n dea. this PERF-go PERF-MP-be God POSS ‘He was on his way to Tamale, the twin did not live to see the sacred day, And this gentleman has now gone to be with God.’ c
c
c
c
c
Nti se wo hwe nipa bi sa a, -ye nipa a So when you look human-being some like-that CM he-be human-being rel anka -re-be-ye kunini, na e-t da bi a e-ye a anka He-PROG-MP-be great-person, and it-reach day some CM it-be PART na ye-n-nim nea Onyankop n nso a-hyehye. and we-NEG-know what God also PERF-plan. ‘When you consider someone like this, he would have been a great person but sometimes we do not know what God has also planned.’ c
c
c
(12)
c
c
In (11), the referent of , the late NUGS president, could not fulfil his desire of arriving in Tamale, due to his untimely death, when he was actually on his way (indicated by the progressive marker re, see Section 3.4). In interpreting the utterance, anka leads one to the assumption that there was a desire which could not be fulfilled. Example (12) also illustrates this potential which cannot be fulfilled, again due to death. c
3.3. Anka in communicating the existence of a past state of affairs The role that anka plays in (3), (repeated below), appears to be quite different from the cases already discussed in Sections 3.1 and 3.2. In combination with other lexical items (na 13 Fofie is one of the sacred days on the calendar of the Akans. It is also the day that twins perform some of their rituals.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1007
i.e., ‘then’ in (3a) and kan no i.e., ‘the beginning’ in (3b)), anka introduces a state of affairs that existed at some point in time, but does no longer do so at the time of utterance. Na is an adverbial that refers to a past time. The uses of na and kan no, respectively, make it easier for the interlocutor to infer that the reference in (3) is to something in the past.
c
(b)
Na anka -kyere ade w kurom. Then anka s/he-teach thing be downtown. ‘S/he used to teach downtown.’ Kan no anka Onyankop n ben fam. Beginning DET anka God near ground. ‘Formerly God used to be close to the earth.’ c
(a)
c
(3)
As the state of affairs described no longer exists at the time of utterance, it is inadequate to translate (3) using the simple past tense, i.e., translating (3a) as she did not teach in town and (3b) as God was not close to the earth. This use of anka is perfectly illustrated by the intervention in (13) by a radio panelist (KA) during a discussion as to whether former heads of state of the Republic of Ghana should be included in the council of state, which is an advisory body to the president. One of the discussants (KP), who agrees to this suggestion, indicates that provision to include former heads of state was made in the previous constitutions of the republic (abolished each time there was a coup d’e´ tat); and even in the present constitution, this provision was made at the draft stage. (13) KP: Na e-w mu, na. . ... Then it-be inside, CONJ. . ... ‘It was part of it, but then. . .. . .’ KA: Na anka e-w mu. Then anka it-be inside ‘It used to be part of it’ KP: Na anka e-w mu, na aban no na e-yi fi-i mu. Then anka it-be inside, and government DET FP it-take come-COMPL inside ‘It used to be part of it, it was the government which took it out.’ c
c c
KP started by merely stating that this provision mentioned earlier on was found in the constitution. However, using anka, KA intervenes to draw the attention of KP and the other participants (and in fact all of the listeners) to the fact that this provision did not remain there. In certain circumstances, this sort of information may be conveyed with na; however, KA’s intervention and KP’s subsequent acceptance indicate that when anka is included, then there is absolutely no doubt that the proposition expressed can no longer be said to be true. As mentioned earlier, utterances of this nature communicate more than a simple negative proposition, as in she does not teach in town and God is not close to the earth ((3a) and (3b), respectively). The relevance of the utterances lies in the fact that a certain state of affairs was true in the past, but that this is no longer the case—which is exactly what anka
1008
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
does in collaboration with na and kan no. As already mentioned, the use of anka in (3) deviates from its uses in (1) and (2), in that the state of affairs being described is not desirable, let alone potential. 3.4. Anka in communicating acts perceived to be very close to happening The fourth use of anka, unlike the first, cannot be interpreted as a politeness marker. Also, this use differs from the use shown in (2) in that it is not found in utterances expressing intentions and desires. Moreover, since there is no indication of the existence of a past state of affairs, as described in these utterances, the use of anka in (4) cannot be said to be identical to its use in (3) either. Admittedly, with regard to this fourth use of anka, one can talk about a semblance of a potential which was not fulfilled, as in (1) and (2). And also, there is an affinity to the examples in (3), to the extent that these utterances are set in the past. Nonetheless, this fourth use warrants categorization in a separate section because of the peculiarity of anka occurring in narrated past events, where there were expectations which were never realised. It is not the case that the events talked about here are states of affairs which existed at some time in the past, but are no more (cf. Section 3.3); nor does the narration refer to propositions that are desirable and or potential (cf. Sections 3.1 and 3.2). The events described are completed, even though the expectations that were raised as a result of perceiving certain things, were not met. What is happening in (4), repeated below, is that in combination with the present progressive aspect (PROG) represented as re, anka indicates that what is described did not actually happen, but was perceived as very close to happening. (4) (a) Ns e yi wo-w bea no ne ne ba ara ma Thorns these RED-prick-COMPL woman DET and her child so-much-so that anka ade re-ye a-t w n so. anka thing PROG-do PERF-fall them top. ‘These thorns pricked the woman and her child so much so that they almost fainted.’ (b) Kwadwo serew no ara ma anka ne mfe re-bu. Kwadwo laugh.COMPL her so-much-so that anka his ribs PROG-break. ‘Kwadwo laughed at her so much so that his sides almost broke.’ c
c
cc
c
c
The use of the present progressive prefix re would usually suggest that this was something which the subject(s) was (were) in the process of doing. The presence of anka, however, suggests that the referents never arrived at the ‘expected end’. For instance in (4a), the thorns pricked the woman and the child to such an extent that anyone looking on expected something horrible (like fainting) to happen to them. Still, even though they probably looked like they were fainting, in the final analysis they did not; however, fainting was what one would have expected them to do under these circumstances. The anka-clause in (4b) anka ne mfe rebu (‘my ribs almost broke’) is a hyperbole. To the writer, that was the best way to describe the extent to which Kwadwo laughed. His ribs did not break, but he laughed so hard that one would not have been surprised if it actually had come to that. The speaker relates his interpretation of some perceived stimuli, which
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1009
(hyperbolically speaking) indicated to him at the time that, due to Kwadwo’s uncontrolled laughter, his ribs broke; but again, by the time of utterance it has become clear to the speaker that, after all, things were not all that bad. The anka-clauses thus express a kind of ‘expectation’ of which the referents fall short; this explains the use of ‘almost’ in the glosses. Even so, I would hesitate to gloss anka by itself as ‘almost’. The ‘almost’ interpretation is made available to the interlocutor as a result of the encoded meaning of anka in its interaction with the present progressive prefix re and other pragmatic information derived from the context, such as what the speaker physically perceived (as in the examples in (4)). This special use of anka is restricted to references to past events described (and indeed interpreted) by the speaker. As mentioned at the beginning of this section, this fourth use of anka is unlike the third use, which also deviates from the first and second uses. Common to all these instances ((1) to (4)) is that what is described by means of the clause containing anka does not exist as a factual state of affairs at the time of utterance. All other meaning components that arguably enter into the pragmatic interpretation of the linguistic structures illustrated above are triggered by the presence of anka (in conjunction with certain other indicators, like the tense aspect marker re in (4), temporal expressions na and kan no in (3), and general sentence structures in (2) and (1)).
4. The meaning of anka Figuring out the meaning of anka is quite a challenge. Christaller (1933: 219) gives us a good starting point. He defines anka as follows: ‘‘This word is put at the head of a sentence to indicate that the idea expressed in the sentence is no present reality, but either (a) a thing of the past, or (b) merely imaginary, its reality being made impossible by another fact. . .’’ This tells us part of the story, but not the whole story. Christaller’s definition covers examples like those in (2) and (3). Other categories like (1) and (4) are not accounted for adequately. To describe the examples in (1) as ideas which are ‘‘merely imaginary’’ is not sufficient since they represent concrete desires of the speakers. Though the propositions described in (4) are in reference to past events, what anka encodes is not the reference to the past, but the fact that the events described were perceived as almost happening. Quite apart from the varied ways in which it is used, a lot of interaction goes on between anka and other lexical items in specific utterances, which work in tandem with anka and help to add contextual information. In addition to what has been suggested that anka encodes, much of its meaning in a given utterance will have to be contextually determined; and the problem remains if there is a principled basis for deciding whether this marker is either lexically ambiguous between its different uses, or else lexically monosemous with a potential to be enriched in accordance with those contextual assumptions which are most accessible to the interactant. In the following, I will suggest some preliminary answers to these questions. In combination with expressions like mepe se me sre (I would like to ask) and contextually derived assumptions, anka in (1) is understood to describe potential future
1010
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
states of affairs which are desirable to the speaker, and thus these utterances are easily interpretable as requests. In (2), the desirable and potential states of affairs are described as being contrary to the facts. In (2a), anka interacts with the na-clause to communicate an unreal but desirable state of affairs; the na-clause gives a reason why the desired state of affairs could not be accomplished. In (2b), enye (be not) co-occurring with the conditional marker se. . .a, presents a counterfactuality that prevents the desire expressed in the ankaclause from being fulfilled. In (3), anka combines with the lexical item na and the expression kan no to present states of affairs which no longer exist. Finally in (4), there is interaction between the present progressive marker re and anka to describe something that did not actually happen, but got very close to happening. Do these four different uses of anka mentioned above make the word lexically ambiguous? It is quite simple to reconcile the uses in (1) and (2), based on the fact that the states of affairs described in both instances are desirable and potential (unless a reason is given for inability of fulfilment), while (3) and (4) appear quite different. I would like to claim that there is a core meaning of anka that runs through all the abovementioned uses. It is that the proposition expressed represents a state of affairs which is not real at the time of utterance; the state of affairs is not factual; rather, it is hypothetical or unreal. This core meaning does not include the information why the state of affairs is unreal at the time of utterance, or how the speaker relates to its non-existence. Since, in fact, the specific meaning conveyed by anka in a particular utterance is heavily context-dependent, the further question may be raised whether the utterance containing anka is (a) an indirect request, (b) an unfulfilled intention or desire, (c) the existence of a state of affairs in the past, or (d) an act which was perceived to be very close to happening. This contextual dependence represents an underdetermination which disappears when the encoded meaning of anka combines with other linguistic expressions as well as with extralinguistic information retrieved from a variety of perceptual and cognitive sources. 5. Anka: a modal marker Based on the preceding, I propose to classify anka as a modal marker. A modal linguistic device relates to mood and modality, which have to do with the speaker’s attitude to the proposition expressed or manner as contrasted with substance. The attitude could be epistemic, including certainty, possibility and doubt, or deontic, having to do with obligation and permission, or it could relate to the speaker’s (or someone else’s) intentions, desires or hopes. In Indo-European languages, such as French, Spanish, and German, this aspect of speaker attitude is usually encoded by means of bound morphemes, like subjunctive suffixes attached to verb forms. This does not appear to be the case in Akan. Akan anka, as we have seen above, plays the role of a marker conveying a modal meaning. It encodes what relevance theory refers to as a higher-level explicature14 which, in relation to a speaker S and a proposition P, conveys that in S’s opinion P is not factual. As 14 A higher level explicature is ‘‘a particular kind of explicature which involves embedding the propositional form of the utterance or one of its constituent propositional forms under a higher-level description such as a speech-act description, a propositional attitude description or some other comment on the embedded proposition’’ Carston (2002: 377).
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1011
mentioned in the previous section, what makes P not factual, and what determines the speaker’s attitude to the fact that P is not factual, are considerations of relevance in the current context. Using this encoded, higher-level explicature as a basis, the interlocutor in specific utterances includes contextual information to make specific inferences. In (1) and (2), the specifically inferred higher-level explicature communicated is (14): (14)
I desire that P.
\
In (3), the positive proposition P (respectively s/he teaches in town, and God is close to earth) is rejected. The inferred higher level explicature thus communicated is (15): (15)
I disbelieve that P.
In (4), the inferred higher level explicature communicated is the interpretation or metarepresentation of (16), projected to some time in the past, with anka triggering the inference that the expectation was not fulfilled. (16)
I expected that P.
In (2), what is further communicated by the modal marker anka is the failure to bring to fulfilment the assumption communicated by P. For example in (2a), the assumption that he will come is understood to be denied. The same goes for (2b) as well as for (2c). The assumptions that she will come and see the child and that he will go are denied. In (3), although the speaker at present believes that P, what is communicated, in addition to that epistemic attitude, is that she believes that P was true at a certain time in the past. Anka occurs clause-initially, except when it co-occurs with other markers (compare the examples in (2c) and (3)), which is when the other markers take the initial position before anka. It takes scope over the clause within which it occurs; however, when this clause is embedded in a conditional, the modal element encoded by anka takes scope over the whole utterance. Consequently, this forces the insertion of the same marker anka in the main clause, as illustrated in (17). (In Akan, conditionality is expressed by a discontinuous expression se. . .a, where se is optional). (17)
Se anka me-di a, anka me-ka a-kyere wo. CM anka I-eat, CM, anka I-say CONS-show you. ‘If I would have liked to eat, I would have told you.’
Another point to consider is what happens when anka is used in the same utterance with a negation, since the encoded meaning of anka has got something to do with negation, more specifically counterfactuality. In (1) and (2), what happens is simple. The communicated propositional attitude (i.e., higher-level explicature) that would be recovered from negated anka clauses is the opposite of what is recovered from the non-negative anka clauses
1012
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
appearing in (1) and (2), though most of the examples (with the possible or probable exception of (2a)), would end up sounding pragmatically odd. What happens when utterances such as (3) are negated is more interesting. In Akan, negation is realised as a homorganic nasal that is prefixed to a verb and is orthographically represented as /n/ or /m/. When (3a) and (3b) are negated the result is as in (18) and (19): (18)
?Na anka -n-kyere ade w kurom. Then anka s/he-NEG-teach thing be town. ‘It is not the case that she used to teach in town.’
(19)
?Kan no anka Oynyankop n m-men fam. Beginning DEF anka God NEG-close ground. ‘It is not the case that in the past God used to be close to the earth.’
c
c
c
As statements, (18) and (19) are grammatically odd. They are only acceptable as negative questions, which give a specific interpretation of S’s thought. In this case, the last syllable is produced on a relatively low pitch; the utterances will translate as (20) and (21). The declarative form represents the speaker’s belief (propositional attitude), while the negative polarity of the proposition expressed by (20) and (21) represents a thought attributed to someone other than the speaker. (20) (21)
She used to teach in town, didn’t she? God used to be close to the earth, didn’t He?
In (4), like in most of the cases in (1) and (2), an attempt to negate the anka-clauses result in pragmatically odd statements. The pragmatically (and sometimes grammatically) infelicitous use of negative with anka is the result of the inherent negation in the semantics of anka. Anka is used in representing a state of affairs which is not real and it is implausible to negate a non-existing state of affairs.
6. Summary In this paper, I have considered the various uses of Akan anka, (a) as a politeness marker, (b) in communicating unfulfilled desires and intentions, (c) in communicating the existence of a state of affairs which no longer exists, and (d) in communicating acts which were thought to have been very close to happening but did not actually happen. I have claimed that anka is not ambiguous between these different uses. Rather, it is vague, so that topdown extralinguistic inference is crucial in arriving at the intended interpretation. To consider the encoded meaning of anka as vague rather than ambiguous is a more economical and psychologically plausible way of accounting for this marker, considering the fact that in all its uses, it encodes the assumption that the proposition expressed by the
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1013
utterance in which it occurs, is not factual. Because anka encodes a higher-level explicature, which conveys that (for speaker S, proposition P) P, in S’s opinion, is not factual, I call it a modal marker.
Acknowledgements I am grateful to Thorstein Fretheim and two anonymous reviewers for their careful reading and critical comments, which greatly improved the quality of this paper. Any remaining shortcomings are mine. An earlier draft of this paper was presented at the Linguistics Association of Great Britain (LAGB) Spring 2000 meeting. I am grateful to that audience as well.
References Amfo, Nana Aba Appiah, 2001. A relevance-theoretic study of some pragmatic markers in Akan. Working Papers in Linguistics, vol. 36. NTNU, Trondheim. Blakemore, Diane, 1987. Semantic Constraints on Relevance. Blackwell, Oxford. Blakemore, Diane, 2000. Indicators and procedures: nevertheless and but. Journal of Linguistics 36, 463–486. Blass, Regina, 1990. Relevance relations in discourse: a study with special reference to Sissala. Cambridge University Press, Cambridge. Christaller, J.G., 1933. A Dictionary of the Asante and Fante Language Called Tshi (Twi), 2nd edition Basel Evangelical Missionary Society, Basel (1st edition, 1881). Carston, Robyn, 2002. Thoughts and Utterances: The Pragmatics of Explicit Communication. Blackwell, Oxford. Clark, William, 1991. Relevance theory and the semantics of non-declaratives. Ph.D. thesis, University of London. Duranti, Alessandro, Goodwin, Charles (Eds.), 1992. Rethinking Context: Language as An Interactive Phenomenon. Cambridge University Press, Cambridge. Goffman, Erving, 1974. Frame Analysis: An Essay on the Organization of Experience. Harper and Row, New York. Grice, H. Paul, 1989. Studies in the Way of Words. Harvard University Press, Cambridge, MA. Obeng, Samuel G., 1999. Requests in Akan Discourse, 41. Anthropological Linguistics, 2. Sperber, Dan, Wilson, Deirdre, 1995. Relevance: Communication and Cognition. 2nd edition. With a Postface. Oxford, Blackwell. (1986). Wilson, Deirdre, Sperber, Dan, 2004. Relevance theory. In: Horn, Laurence R., Ward, Gregory (Eds.), The Handbook of Pragmatics. Blackwell, Oxford & Boston.
Nana Aba Appiah Amfo is a lecturer at the Department of Linguistics, University of Ghana and a PhD candidate at the Norwegian University of Science and Technology (NTNU). Her current research interests include the use and relevance pragmatic markers in communication, lexical pragmatics and the general use of language in society.
Modal marking in Akan: the case of anka Nana Aba Appiah Amfo Department of Language and Communication Studies, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway Received 6 June 2002; received in revised form 2 December 2004; accepted 14 December 2004
Abstract The current article considers a marker, anka, in Akan (Niger-Congo, Kwa branch), which has a variety of uses. Adopting a relevance-theoretic approach, I argue that anka is a marker with a modal meaning. In spite of the variety of uses of this word, I propose a univocal lexical semantics for it, which will be shown to account for its context-dependent function in all the categories. Taking into consideration the interplay between the encoded meaning of anka and pragmatically derived information in actual speech situations, monosemy is seen as a more economical and psychologically plausible way of accounting for the communicative function of this Akan marker. # 2004 Elsevier B.V. All rights reserved. Keywords: Encoded meaning; Higher-level explicature; Implicature; Modality; Relevance; Ambiguity; Vagueness; Indirect requests; Akan
1. Introduction The focus of this paper is the Akan marker anka. This marker has four disparate uses. Using the framework of relevance theory, my aim in this paper is to look at how closely
Abbreviations: CM, Conditional marker; COMP, Complementizer; COMPL, Completive aspect; CONS, Consecutive; DET, Determiner; FP, Focus particle; FUT, Future tense; H, Hearer/Interlocuteor; IM, Interpretive use marker; MP, Motional prefix; NEG, Negation marker; P, Proposition; PART, Particle; PERF, Perfective aspect; POSS, Possessive pronoun; PROG, Progressive aspect; RED, Reduplication; S, Speaker E-mail address: [email protected]. 0378-2166/$ – see front matter # 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.pragma.2004.12.008
998
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
related these different uses are and if there is a possibility of a form of unification, taking into consideration the fact that ‘‘senses are not to be multiplied beyond necessity’’ (Modified Occam’s Razor: Grice, 1989: 47). First I consider the various ways in which anka is used: (a) as a politeness marker in the performance of indirect requests, (b) as a marker in communicating someone’s unfulfilled intentions and desires in what serves as a kind of counterfactual conditional construction, (c) as a marker communicating the existence of a state of affairs which belongs to the past and has ceased to exist, and finally (d) as a marker which indicates that what is described did not actually happen but almost happened. My next concern is with the lexical meaning that anka encodes. My claim is that anka is vague rather than ambiguous, i.e., it has a univocal lexical meaning, which serves as a basis for context-dependent pragmatic enrichment in various directions. In all its uses it indicates that the proposition expressed represents a state of affairs that is not in existence at the time of utterance. The specific meaning of anka in a given utterance is a function of the interaction of its core meaning with other semantically encoded information as well as pragmatically derived information, which the principle of relevance (Sperber and Wilson, 1995: 271) makes readily available to the recipient. Indeed I do acknowledge the fact that anka interacts regularly with other items in specific utterances to bring out its particular pragmatic relevance in these utterances; however, an attempt to give an account of all such items which anka interacts with goes beyond the scope of this paper and will detract attention from the main focus of this paper.1 My claim is that anka is a modal marker, because it encodes a specific epistemic attitude to the proposition expressed by the clause in which it appears. More particularly, it encodes the assumption that the state of affairs described by the propositional content is not factual at the time of the utterance.
2. Cognitive effects and procedural meaning The analysis made in this paper is couched in the cognitively based inferential theory of communication, called Relevance Theory (Sperber and Wilson, 1986, 1995). According to this framework, an utterance is relevant when it combines with contextual assumptions to yield contextual effects. Such effects can be achieved in three ways. First, a new piece of information is relevant when it combines with existing assumptions to derive contextual implications. Second, a new piece of information may strengthen an existing assumption. Finally, a new piece of information may be relevant when it is inconsistent with an existing assumption and as a result one of the pieces of information will have to be dropped in favour of the other. Utterances are interpreted within specific contexts. Context is quite a contentious concept to deal with in the field of language and communication studies. It is defined differently depending on the perspective taken by a scholar. Duranti and Goodwin, in the 1 However, the contributions of some of these accompanying markers, like se, (see Section 3.2) and na, (Section 3.3) are mentioned, where relevant.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
999
introduction to their edited volume, Rethinking Context: language as an interactive phenomenon, recognize the probable impossibility of providing a ‘‘single, precise and technical definition of context’’ (Duranti and Goodwin, 1992: 2). They admit that researchers need not agree on a single definition of context, as different notions of context may be appropriate to different levels of organization. The chapters of this book are significant, in that there is a demonstration of a shift from the notion of context as a frame that surrounds talk (Goffman, 1974) to one in which talk constitutes a main source for the organization of context. Context according to Sperber and Wilson is psychologically defined as ‘‘a subset of the hearer’s assumptions about the world’’2 (Sperber and Wilson, 1995: 15). This apparent simplicity of their notion of context is an attempt not to favour a variable of context over another. For them, all the ethnographic, social, cultural and participant variables which affect context do so by shaping the assumptions that the interlocutor3 uses in the interpretation of an utterance. In a particular interpretation process, the set of assumptions used is the context of the utterance. But how is the context for an utterance chosen by the interlocutor? Certain linguistic indicators in an utterance direct the interlocutor to a set of assumptions that he brings to bear in the interpretation process.4 Linguistic items of this nature are known in relevance-theoretic terms as encoders of procedural meaning. Being in the minority, such linguistic items are contrasted with the majority set of words and expressions, which encode concepts which the interlocutor utilises in constructing the truth-conditional content of the utterance. This relevance-theoretic procedural/conceptual distinction was introduced by Blakemore (1987) with her re-analysis of some of Grice’s conventional implicatures as semantic constraints on implicatures. Further research (Amfo, 2001; Blakemore, 2000; Blass, 1990; Wilson and Sperber, 2004) has revealed that markers of procedural meaning do not only aid the interpretation process by constraining comprehension at the implicit level; they have the potential of influencing comprehension even explicitly.5 The Akan marker anka that I consider in this paper encodes a procedure by which the utterance which contains anka is to be considered as representing a state of affairs which does not exist at the time of utterance. This state of non-existence becomes part of the assumptions constraining the interpretation of the utterance.
2 The ‘‘hearer’s assumption about the world’’ includes all the pieces of information brought into the interpretation process, including what the hearer believes are the assumptions the speaker had in mind in the production of a particular utterance (cf. Sperber and Wilson, 1995: 15–16.). The goal of a pragmatic theory must be to account for utterance comprehension, a ‘mind-reading’ process based on whatever evidence the speaker’s stimulus gives the interlocutor to derive contextual effects from it. 3 Following from the presumption of optimal relevance (Wilson and Sperber, 2004: 612), the speaker in producing an utterance takes into consideration what she believes to be the assumptions held by the interlocutor via the context. 4 For a more detailed discussion on context as a psychological construct, see Sperber and Wilson, 1995 (1986). 5 Grice took comprehension a step further from simply encoding and decoding by distinguishing between ‘what is said’ and ‘what is meant’ (implicatures). However, his distinction also differs quite significantly from the relevance-theoretic explicit/implicit distinction. For more on this, see Carston, 2002.
1000
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
3. Uses of anka Anka is used in a variety of ways. The examples below illustrate four main areas where it is found:6 1. Politeness: (a) Mo-a-w fufuo? Anka me-di o. You-have-pound fufu? Anka I-eat o. ‘Have you pounded fufu? I would have liked to eat.’ (b) Anka me-pe se me-sre w’akondwa yi kakra. Anka I-like COMP I-beg your’chair this little. ‘I would have liked to borrow your chair for a while.’ (Could I borrow your chair please?) c
2. Unfulfilled intentions and desires: (a) Anka -re-ba ha na -te-e se Kofi n-ni h . Anka he-PROG-come here CONJ he-heard-COMPL COMP Kofi neg-is there. ‘He was coming here, and then he heard Kofi was not here.’ (b) Enye se m’a-nya kataa a, anka me-ba a-be-hwe abofra no. It-NEG-be CM I’PERF-get catarrh CM, anka I-come CONS-MP-look child DEF. ‘Had it not been that I have caught a cold, I would have come to see the baby.’ (c) Se anka e-w se wo-tumi k bi. IM anka it-be COMP you-be.able go some. ‘You know, you should have been able to go along.’ c
c
c
c
c
3. A past state of affairs: (a) Na anka -kyere ade w kurom. Then anka s/he-teach thing be downtown. ‘S/he used to teach downtown.’ (b) Kan no anka Onyankop n ben fam.7 Beginning DET anka God near ground. ‘Formerly God used to be close to the earth.’ c
c
c
4. Acts perceived as almost happening: (a) Ns e yi wow bea no CONJ ne ba ara ma Thorns these RED.prick-COMPL woman DET and her child so-much-so that anka ade re-ye a-t w n so. anka thing PROG-be PERF-fall them top. ‘These thorns pricked the woman and her child so much so that they almost fainted.’ c
c
c
7
c
cc
6
Fufu is a type of West African dish prepared by pounding boiled cassava, yam, plantain or cocoyam. This example is taken from Christaller’s (1933) dictionary.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1001
(b) Kwadwo serew no ara ma anka ne mfe re-bu. Kwadwo laugh.PST her so-much-so that anka his ribs PROG-broke ‘Kwadwo laughed at her so much so that his sides almost broke.’8 3.1. Anka as a politeness marker in indirect requests Obeng (1999)9 categorises Akan requests into direct and indirect ones. Indirect requests can further be either conventional or non-conventional. For conventional indirect requests, Weizman (1989), as quoted in Obeng (ibid.), notes that in English, the speaker makes use of some grammatical or semantic device used conventionally for the purpose of requesting; such devices guide the interlocutor in the recovery process. Among the devices are (a) the grammatical structure of questions with modals such as ‘can’ or ‘could’; and (b) the semantic meaning of apologetic expressions such as ‘please’. Such grammatical or semantic devices to signal the indirectness of a request are found in a number of languages, and my argument in this section is that Akan anka should be considered as such a conventional indirect request marking device; anka is a sort of hedge which tones down an otherwise direct request. Anka lexically encodes the procedure that the proposition encoded by the utterance in which it is contained should be considered as unreal. However, the examples given in (1) are only pragmatically appropriate as indirect requests. The request interpretation is not due to the presence of anka; anka is a semantic device used to mitigate the directness of the request. In (1a), if the speaker had simply said medi o (I will eat), it would still be understood as a request but in this case a direct one.10 (1a) thus serves as an indirect request as a result of the presence of anka. In (1b), the semantic content of the verbs used (that is pe (like) and sre (beg)) contributes to the indirectness, with the presence of anka giving the indirectness a further boost. A characteristic of pragmatic markers is that they very often work in tandem with other linguistic items in an utterance to achieve the desired results; the fact that they are highly context sensitive accounts for the apparent difficulty in their semantic categorisation. There are several ways of achieving indirectness and in (1) it is done by the speaker presenting the situation described as nonexistent due to the lexical meaning of anka. Since in Akan, indirection is one of the ways in which politeness finds expression, anka used in such instances can thus be labelled as a politeness marker. The pragmatically inferred meaning of anka in these examples is reminiscent of the way that Sperber and Wilson (1995) and Clark (1991) handle imperatives. According to these authors, imperatives describe a state of affairs, which is both potential and desirable (to someone, frequently the speaker). In (1a), eating fufu is described as desirable to the speaker, but the proposition expressed describes a state of affairs which is only potential, 8 The examples in (4) are taken from a collection of fables in Akuapem Twi ‘Ananse Akuamoa (1961, Bureau of Ghana Languages). The rest of the examples are in Asante Twi. 9 For a detailed discussion on requests in Akan, see Obeng (1999). 10 Obeng (1999) states that direct requests in Akan are in the form of command sentences, that is, imperatives. This category can, however, be expanded to include declaratives which state the speaker’s desire. Note that I define a direct request as a request performed by virtue of the meaning that the sentence encodes, while the term ‘indirect request’ is reserved for cases where understanding of the illocutionary act as a request depends crucially on the inferential processing informed by the encoded logical form of the sentence.
1002
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
not factual. Even though the sentence without anka still describes a non-factual state of affairs, the speaker’s use of anka is significant in that it signals that he cannot take it for granted that the state of affairs described by the proposition will exist, even after the utterance has been uttered. The speaker’s propositional attitude can be inferred to be one of desire; it also comprises the hope that eating fufu is a potential state of affairs which could be brought into reality by the interlocutor acting in such a way that fruition is accomplished. This process seems to be outside the speaker’s own control, and it is therefore up to the interlocutor to decide what to do in this situation. The utterance is understood as a request, and the use of anka makes a considerable contribution to the polite nature of this request. The state of affairs described in (1b) is likewise unreal in the sense of unfulfilled (but at the same time both potential and desirable). Taken literally, these utterances seem to say that my intention is or was for P to come true, where P unfortunately represents a state of affairs which may not come into existence. However, a native interlocutor will not stop drawing inferences at this point. Every native speaker that I presented with a form similar to those in (1), such as anka mepe se me yusu wo fon no (I had wanted to use your phone) or anka merefa wo pen yi (I had wanted to take your pen), either immediately made an offer (such as ‘here you are’, ‘okay’), or declined, giving an explanation. According to a relevance-theoretic interpretation, an assumption is relevant when it achieves contextual effects in a given context (Sperber and Wilson, 1995:112); however, achieving the required contextual effects demands a processing effort. The smaller the processing effort, the more relevant the assumption, and an added function word like anka, in what is already a grammatical sentence, means added processing cost. Why, then, will a speaker of Akan use anka, when its relevance in the utterance requires extra processing effort, as compared to simply making a direct request (without anka)? The reason is that using anka yields certain contextual effects which would not be achieved if the speaker had chosen to use a linguistic form which simply encodes the assumption of being polite, as does the phrase mepa wo kyew (please). This is a set phrase whose conventional meaning does not allow for much variation in its impact on the comprehension process with regard to context-based factors. In contrast, the inclusion of anka in (1) makes manifest a set of assumptions forming part of the context within which the utterance is expected to be processed. These assumptions may include one or more of the following: (i) (ii) (iii)
S realises that the request is not something that H will normally expect from S. S suspects H may not yield easily due to (i) or for some other reason which will have to be deduced contextually. S acknowledges that S and H are in an (extra-linguistic) environment in which H’s right to exert an influence on the practical consequences of the verbal interaction overrides S’s right to do so.
According to Obeng (ibid.), in Akan society requests are usually not considered impositions on the requestees, due to the society’s collective culture and social interdependence. For this reason, certain (direct) requests, which would be deemed impolite and probably inappropriate, even face-threatening in other societies, are generally not perceived as such in an Akan context. By extension, requests in Akan society, as compared to certain Western
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1003
societies, are non-imposing linguistic acts. This holds true to a large extent for rural Akan communities, where a high level of dependence among members is to be expected. In urban and semi-urban Akan communities (such as Koforidua, where the data in (1) was collected), social interdependence is gradually on the decline and as a result, the way the Akan conceptualise requests may vary, depending on certain socio-pragmatic circumstances. But also here, as Obeng is quick to add (ibid.), the interpersonal relationship between the requester and the requestee should not be overlooked. The requester in (1) certainly ignores his interpersonal relationship with the requestee as well as other socio-pragmatic considerations. An able-bodied person who asks for food (a very basic need) from someone who may only qualify as a mere acquaintance certainly threatens the interlocutor’s negative face. And note that the request was not made because the requester saw the requestee eating or being about to eat; it was made simply on the assumption that the requestee had prepared food. Now, let us see how the above assumptions work with respect to example (1a). This utterance was said by a young man (Y) to a 19-year-old girl (E). They had met once or twice in the past, with minimal interaction having occurred between them. I reproduce (1a) below, including E’s reply as (5). Y began by asking E if her parents were home and then continued: Y: Mo-a-w fufuo? Anka me-di o. You-PERF-pound fufu? Anka I-eat o. ‘Have you pounded fufu? I would have liked to eat.’ E: Eneese wo-a-nya, yen na ye-noa aduane ma wo-di. Then COMP you-PERF-get, we FM we-cook food give you-eat. ‘Oh is that what you think? Is it we who cook for you?’ (Is it our responsibility to cook for you?) c
(5)
Given the not-so-close interpersonal relationship between Y and E, a request for food in the above-mentioned circumstances is not something that E would normally expect from Y, and Y is probably aware that, as a result of this, E will not accede to his request, especially if it is made directly. Even a request that is toned down by the use of anka would, in this particular case, not only come unexpected for E, but appear inappropriate as well, something which her reply confirms. In using anka, Y implies that he is aware of all these assumptions, which would normally have prevented him from making such a request directly. However, since he prefers to disregard the socio-cultural assumptions in order to have his desire fulfilled, he makes the awareness of the social inappropriateness of his request manifest, and hopes that in doing so, he will cause E to yield to his request. Presumably, he felt that this might be the only way to make E recognise his communicative intentions and at the same time give himself a chance to achieve his goal. The use of anka as a politeness marker is further illustrated by the following examples from a radio discussion. This discussion took place during a review programme. The newspaper article under review was supposed to have been based on a Serious Fraud Office (SFO) interim report, which apparently was not in favour of some public official. The report in question had not actually been made public officially, but the newspaper which published the article claimed to have access to it anyway. In (6), one of the panelists appeals to the relevant authorities to make such reports public in order to put an end to speculations,
1004
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
while another panelist in (7) reiterates an official’s request that journalists should exercise some restraint when dealing with reports which are not considered as final.11 Anka me-ka se se SFO ye rep t bi a they should make the Anka I-say COMP CM SFO do report some CM. . .. . . report open to the public. ‘I would request that when the SFO prepares a report, they should make it c
(6)
open to the public.’ Anka -be-sre se anka nsentwerefo ye-n-to yen bo Anka he-FUT-beg COMP anka newswriters we-IMP-throw our chest ansa, na SFO na ew finale. . .e. . . mandate or power before, CONJ SFO FM it-has final mandate or power ‘He would ask that journalists exercise restraint because it is the SFO which has the final mandate (in such matters).’ c
c
(7)
c
In both (6) and (7), the speakers use anka as a downtoning marker while making their requests, since they know that unless the interlocutors act as suggested, there is nothing that they can do practically about these situations. In (7), note that anka is used both in the main clause and in the complement clause. The first anka used in the utterance is an indication that the speaker is making an indirect request, that is, for journalists to take their time in matters of that nature. The second anka directs the interlocutor to a particular implicature,12—that journalists are not taking their time. In other words, the proposition that journalists should take their time (which is the message explicated without anka), is not true at the time of utterance. This particular implicature becomes part of the context within which the utterance is interpreted. This particular usage of anka, where the speaker states what will be desirable if the pragmatic conditions are or (in counterfactual situations) were fulfilled, is comparable to a familiar pattern of the usage in English, where a potentially face-threatening act is prefaced with an expression involving the past tense form of modals. (8)
I would have liked to ask you. . .
The speaker will often times go ahead with her (indirect) question or request in spite of her apparently counterfactual preface, having first expressed her realisation that the question or request might be deemed unsuitable under the circumstances. An utterance like (9), produced by a supervisor addressing a student, would normally be relevant beyond what seems to be literally an example of wishful thinking. (9)
It would have been a good thing if you and I could get together to discuss your paper.
11 It is not uncommon to find code mixing and code switching among educated Akans and I produce the examples as they were recorded. 12 An implicature, according to Carston (2002: 377), is a ‘‘communicated assumption which is derived solely via processes of pragmatic inference’’.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1005
3.2. Anka in communicating unfulfilled intentions and desires Another use of anka is found in utterances expressing intentions or desires which cannot be, or have not been, fulfilled (as exemplified in (2), repeated here for convenience). The reason for the unfulfilled intention or desire is often, though not always, given in a conjoined clause (as in (2a)) or a conditional clause (as in (2b)). The anka-clauses in (2), unlike those in (1), are not interpreted as requests or entreaties, due to the general sentence constructions. The forms in (2) are counterfactual conditionals with the protasis introduced by anka. However as in (1), they describe a desirable and a potential state of affairs. (2) (a) Anka -re-ba ha na -te-e se Kofi n-ni h . Anka he-PROG-come here CONJ he-hear-PST COMP Kofi NEG-is there. ‘He was coming here, but then he heard Kofi was not here.’ (b) E-n-ye se m’a-nya kataa a, anka me-ba a-be-hwe abofra no. It-NEG-be CM I’PERF-get catarrh CM anka I-come CONS-MP-look child DEF. ‘Had it not been that I have caught a cold, I would have come to see the baby.’ (c) Se anka e-w se wo-tumi k bi. IM anka it-be COMP you-be.able go some. ‘You know, you should have been able to go along. c
c
c
c
c
In (2a), the clause expressing the intention starts the utterance, before an explanation is given for the failure to carry out the action. The anka-clause describes a desirable and potential state of affairs, or at least so it seemed to the referent of the pronominal (he), until in the conjoined clause, he was informed of Kofi’s absence, which went against what was implicitly assumed in the protasis. The relevance of the anka-construction is that while it expresses a potential and desirable state of affairs, it at the same time implicates that the actual performance of this state of affairs is hindered by something. What anka does is to make more manifest a specific implicature of (2a). As the state of affairs referred to in the clause modified by anka does not exist at the time of utterance, and as the state of affairs described would have belonged to the future in any case, the inferred meaning is that the referent of (he) is not coming. In (2b), the reason for the inability to carry out the intention is given before the intention is expressed. The situation described in the main clause is desirable, and it would also have been potential if it were not for the fact that the negative proposition expressed in the preceding conditional clause is presented as counterfactual (due to the co-occurrence of enye ‘it is not’, and the conditional marker se. . .a). In (2c), the assumption expressed is that the potential and desirable state of affairs described by that utterance is indeed desirable to the speaker, and is furthermore presented by the speaker as being probably desirable to the interlocutor as well. This is in spite of some apparent contrary evidence, which the interactants may have taken to imply that the state of affairs cannot be fulfilled. (2c) was said by a wife to a husband concerning a men’s camp meeting organised by their church. This meeting has not yet taken place, but both the speaker (wife) and the interlocutor (husband) are aware that the interlocutor does not intend c
c
1006
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
to go; this explains the use of the first se in the utterance. This se is not a complementiser (as is the second se in this utterance). It is a marker which indicates that the information that I’m about to share is not new information, and you know that as well as I do. What is explicitly communicated in (2c) is that the interlocutor should be able to go. The combination of se and anka guides the interlocutor to the assumption in (10), which forms part of the context within which (2c) is processed: (10)
S knows that I know that she is aware I am not going, but it is desirable (to her) that I go, and both of us know that.
When anka is used in these instances, it gives rise to the contextual assumptions that some desire or potential cannot be achieved. The following two examples, said of a president of the National Union of Ghana Students (NUGS) who lost his life in a car accident just a month after his inauguration. again illustrates the use of anka in unfulfilled desires and potentials. (11) Anka -re-k Tamale a fofie13 an-to Atta na abrantie Anka he-PROG-go Tamale PART sacred-day NEG-arrive twin CONJ gentleman yi a-k a-k -ye Nyankop n dea. this PERF-go PERF-MP-be God POSS ‘He was on his way to Tamale, the twin did not live to see the sacred day, And this gentleman has now gone to be with God.’ c
c
c
c
c
Nti se wo hwe nipa bi sa a, -ye nipa a So when you look human-being some like-that CM he-be human-being rel anka -re-be-ye kunini, na e-t da bi a e-ye a anka He-PROG-MP-be great-person, and it-reach day some CM it-be PART na ye-n-nim nea Onyankop n nso a-hyehye. and we-NEG-know what God also PERF-plan. ‘When you consider someone like this, he would have been a great person but sometimes we do not know what God has also planned.’ c
c
c
(12)
c
c
In (11), the referent of , the late NUGS president, could not fulfil his desire of arriving in Tamale, due to his untimely death, when he was actually on his way (indicated by the progressive marker re, see Section 3.4). In interpreting the utterance, anka leads one to the assumption that there was a desire which could not be fulfilled. Example (12) also illustrates this potential which cannot be fulfilled, again due to death. c
3.3. Anka in communicating the existence of a past state of affairs The role that anka plays in (3), (repeated below), appears to be quite different from the cases already discussed in Sections 3.1 and 3.2. In combination with other lexical items (na 13 Fofie is one of the sacred days on the calendar of the Akans. It is also the day that twins perform some of their rituals.
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1007
i.e., ‘then’ in (3a) and kan no i.e., ‘the beginning’ in (3b)), anka introduces a state of affairs that existed at some point in time, but does no longer do so at the time of utterance. Na is an adverbial that refers to a past time. The uses of na and kan no, respectively, make it easier for the interlocutor to infer that the reference in (3) is to something in the past.
c
(b)
Na anka -kyere ade w kurom. Then anka s/he-teach thing be downtown. ‘S/he used to teach downtown.’ Kan no anka Onyankop n ben fam. Beginning DET anka God near ground. ‘Formerly God used to be close to the earth.’ c
(a)
c
(3)
As the state of affairs described no longer exists at the time of utterance, it is inadequate to translate (3) using the simple past tense, i.e., translating (3a) as she did not teach in town and (3b) as God was not close to the earth. This use of anka is perfectly illustrated by the intervention in (13) by a radio panelist (KA) during a discussion as to whether former heads of state of the Republic of Ghana should be included in the council of state, which is an advisory body to the president. One of the discussants (KP), who agrees to this suggestion, indicates that provision to include former heads of state was made in the previous constitutions of the republic (abolished each time there was a coup d’e´ tat); and even in the present constitution, this provision was made at the draft stage. (13) KP: Na e-w mu, na. . ... Then it-be inside, CONJ. . ... ‘It was part of it, but then. . .. . .’ KA: Na anka e-w mu. Then anka it-be inside ‘It used to be part of it’ KP: Na anka e-w mu, na aban no na e-yi fi-i mu. Then anka it-be inside, and government DET FP it-take come-COMPL inside ‘It used to be part of it, it was the government which took it out.’ c
c c
KP started by merely stating that this provision mentioned earlier on was found in the constitution. However, using anka, KA intervenes to draw the attention of KP and the other participants (and in fact all of the listeners) to the fact that this provision did not remain there. In certain circumstances, this sort of information may be conveyed with na; however, KA’s intervention and KP’s subsequent acceptance indicate that when anka is included, then there is absolutely no doubt that the proposition expressed can no longer be said to be true. As mentioned earlier, utterances of this nature communicate more than a simple negative proposition, as in she does not teach in town and God is not close to the earth ((3a) and (3b), respectively). The relevance of the utterances lies in the fact that a certain state of affairs was true in the past, but that this is no longer the case—which is exactly what anka
1008
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
does in collaboration with na and kan no. As already mentioned, the use of anka in (3) deviates from its uses in (1) and (2), in that the state of affairs being described is not desirable, let alone potential. 3.4. Anka in communicating acts perceived to be very close to happening The fourth use of anka, unlike the first, cannot be interpreted as a politeness marker. Also, this use differs from the use shown in (2) in that it is not found in utterances expressing intentions and desires. Moreover, since there is no indication of the existence of a past state of affairs, as described in these utterances, the use of anka in (4) cannot be said to be identical to its use in (3) either. Admittedly, with regard to this fourth use of anka, one can talk about a semblance of a potential which was not fulfilled, as in (1) and (2). And also, there is an affinity to the examples in (3), to the extent that these utterances are set in the past. Nonetheless, this fourth use warrants categorization in a separate section because of the peculiarity of anka occurring in narrated past events, where there were expectations which were never realised. It is not the case that the events talked about here are states of affairs which existed at some time in the past, but are no more (cf. Section 3.3); nor does the narration refer to propositions that are desirable and or potential (cf. Sections 3.1 and 3.2). The events described are completed, even though the expectations that were raised as a result of perceiving certain things, were not met. What is happening in (4), repeated below, is that in combination with the present progressive aspect (PROG) represented as re, anka indicates that what is described did not actually happen, but was perceived as very close to happening. (4) (a) Ns e yi wo-w bea no ne ne ba ara ma Thorns these RED-prick-COMPL woman DET and her child so-much-so that anka ade re-ye a-t w n so. anka thing PROG-do PERF-fall them top. ‘These thorns pricked the woman and her child so much so that they almost fainted.’ (b) Kwadwo serew no ara ma anka ne mfe re-bu. Kwadwo laugh.COMPL her so-much-so that anka his ribs PROG-break. ‘Kwadwo laughed at her so much so that his sides almost broke.’ c
c
cc
c
c
The use of the present progressive prefix re would usually suggest that this was something which the subject(s) was (were) in the process of doing. The presence of anka, however, suggests that the referents never arrived at the ‘expected end’. For instance in (4a), the thorns pricked the woman and the child to such an extent that anyone looking on expected something horrible (like fainting) to happen to them. Still, even though they probably looked like they were fainting, in the final analysis they did not; however, fainting was what one would have expected them to do under these circumstances. The anka-clause in (4b) anka ne mfe rebu (‘my ribs almost broke’) is a hyperbole. To the writer, that was the best way to describe the extent to which Kwadwo laughed. His ribs did not break, but he laughed so hard that one would not have been surprised if it actually had come to that. The speaker relates his interpretation of some perceived stimuli, which
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1009
(hyperbolically speaking) indicated to him at the time that, due to Kwadwo’s uncontrolled laughter, his ribs broke; but again, by the time of utterance it has become clear to the speaker that, after all, things were not all that bad. The anka-clauses thus express a kind of ‘expectation’ of which the referents fall short; this explains the use of ‘almost’ in the glosses. Even so, I would hesitate to gloss anka by itself as ‘almost’. The ‘almost’ interpretation is made available to the interlocutor as a result of the encoded meaning of anka in its interaction with the present progressive prefix re and other pragmatic information derived from the context, such as what the speaker physically perceived (as in the examples in (4)). This special use of anka is restricted to references to past events described (and indeed interpreted) by the speaker. As mentioned at the beginning of this section, this fourth use of anka is unlike the third use, which also deviates from the first and second uses. Common to all these instances ((1) to (4)) is that what is described by means of the clause containing anka does not exist as a factual state of affairs at the time of utterance. All other meaning components that arguably enter into the pragmatic interpretation of the linguistic structures illustrated above are triggered by the presence of anka (in conjunction with certain other indicators, like the tense aspect marker re in (4), temporal expressions na and kan no in (3), and general sentence structures in (2) and (1)).
4. The meaning of anka Figuring out the meaning of anka is quite a challenge. Christaller (1933: 219) gives us a good starting point. He defines anka as follows: ‘‘This word is put at the head of a sentence to indicate that the idea expressed in the sentence is no present reality, but either (a) a thing of the past, or (b) merely imaginary, its reality being made impossible by another fact. . .’’ This tells us part of the story, but not the whole story. Christaller’s definition covers examples like those in (2) and (3). Other categories like (1) and (4) are not accounted for adequately. To describe the examples in (1) as ideas which are ‘‘merely imaginary’’ is not sufficient since they represent concrete desires of the speakers. Though the propositions described in (4) are in reference to past events, what anka encodes is not the reference to the past, but the fact that the events described were perceived as almost happening. Quite apart from the varied ways in which it is used, a lot of interaction goes on between anka and other lexical items in specific utterances, which work in tandem with anka and help to add contextual information. In addition to what has been suggested that anka encodes, much of its meaning in a given utterance will have to be contextually determined; and the problem remains if there is a principled basis for deciding whether this marker is either lexically ambiguous between its different uses, or else lexically monosemous with a potential to be enriched in accordance with those contextual assumptions which are most accessible to the interactant. In the following, I will suggest some preliminary answers to these questions. In combination with expressions like mepe se me sre (I would like to ask) and contextually derived assumptions, anka in (1) is understood to describe potential future
1010
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
states of affairs which are desirable to the speaker, and thus these utterances are easily interpretable as requests. In (2), the desirable and potential states of affairs are described as being contrary to the facts. In (2a), anka interacts with the na-clause to communicate an unreal but desirable state of affairs; the na-clause gives a reason why the desired state of affairs could not be accomplished. In (2b), enye (be not) co-occurring with the conditional marker se. . .a, presents a counterfactuality that prevents the desire expressed in the ankaclause from being fulfilled. In (3), anka combines with the lexical item na and the expression kan no to present states of affairs which no longer exist. Finally in (4), there is interaction between the present progressive marker re and anka to describe something that did not actually happen, but got very close to happening. Do these four different uses of anka mentioned above make the word lexically ambiguous? It is quite simple to reconcile the uses in (1) and (2), based on the fact that the states of affairs described in both instances are desirable and potential (unless a reason is given for inability of fulfilment), while (3) and (4) appear quite different. I would like to claim that there is a core meaning of anka that runs through all the abovementioned uses. It is that the proposition expressed represents a state of affairs which is not real at the time of utterance; the state of affairs is not factual; rather, it is hypothetical or unreal. This core meaning does not include the information why the state of affairs is unreal at the time of utterance, or how the speaker relates to its non-existence. Since, in fact, the specific meaning conveyed by anka in a particular utterance is heavily context-dependent, the further question may be raised whether the utterance containing anka is (a) an indirect request, (b) an unfulfilled intention or desire, (c) the existence of a state of affairs in the past, or (d) an act which was perceived to be very close to happening. This contextual dependence represents an underdetermination which disappears when the encoded meaning of anka combines with other linguistic expressions as well as with extralinguistic information retrieved from a variety of perceptual and cognitive sources. 5. Anka: a modal marker Based on the preceding, I propose to classify anka as a modal marker. A modal linguistic device relates to mood and modality, which have to do with the speaker’s attitude to the proposition expressed or manner as contrasted with substance. The attitude could be epistemic, including certainty, possibility and doubt, or deontic, having to do with obligation and permission, or it could relate to the speaker’s (or someone else’s) intentions, desires or hopes. In Indo-European languages, such as French, Spanish, and German, this aspect of speaker attitude is usually encoded by means of bound morphemes, like subjunctive suffixes attached to verb forms. This does not appear to be the case in Akan. Akan anka, as we have seen above, plays the role of a marker conveying a modal meaning. It encodes what relevance theory refers to as a higher-level explicature14 which, in relation to a speaker S and a proposition P, conveys that in S’s opinion P is not factual. As 14 A higher level explicature is ‘‘a particular kind of explicature which involves embedding the propositional form of the utterance or one of its constituent propositional forms under a higher-level description such as a speech-act description, a propositional attitude description or some other comment on the embedded proposition’’ Carston (2002: 377).
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1011
mentioned in the previous section, what makes P not factual, and what determines the speaker’s attitude to the fact that P is not factual, are considerations of relevance in the current context. Using this encoded, higher-level explicature as a basis, the interlocutor in specific utterances includes contextual information to make specific inferences. In (1) and (2), the specifically inferred higher-level explicature communicated is (14): (14)
I desire that P.
\
In (3), the positive proposition P (respectively s/he teaches in town, and God is close to earth) is rejected. The inferred higher level explicature thus communicated is (15): (15)
I disbelieve that P.
In (4), the inferred higher level explicature communicated is the interpretation or metarepresentation of (16), projected to some time in the past, with anka triggering the inference that the expectation was not fulfilled. (16)
I expected that P.
In (2), what is further communicated by the modal marker anka is the failure to bring to fulfilment the assumption communicated by P. For example in (2a), the assumption that he will come is understood to be denied. The same goes for (2b) as well as for (2c). The assumptions that she will come and see the child and that he will go are denied. In (3), although the speaker at present believes that P, what is communicated, in addition to that epistemic attitude, is that she believes that P was true at a certain time in the past. Anka occurs clause-initially, except when it co-occurs with other markers (compare the examples in (2c) and (3)), which is when the other markers take the initial position before anka. It takes scope over the clause within which it occurs; however, when this clause is embedded in a conditional, the modal element encoded by anka takes scope over the whole utterance. Consequently, this forces the insertion of the same marker anka in the main clause, as illustrated in (17). (In Akan, conditionality is expressed by a discontinuous expression se. . .a, where se is optional). (17)
Se anka me-di a, anka me-ka a-kyere wo. CM anka I-eat, CM, anka I-say CONS-show you. ‘If I would have liked to eat, I would have told you.’
Another point to consider is what happens when anka is used in the same utterance with a negation, since the encoded meaning of anka has got something to do with negation, more specifically counterfactuality. In (1) and (2), what happens is simple. The communicated propositional attitude (i.e., higher-level explicature) that would be recovered from negated anka clauses is the opposite of what is recovered from the non-negative anka clauses
1012
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
appearing in (1) and (2), though most of the examples (with the possible or probable exception of (2a)), would end up sounding pragmatically odd. What happens when utterances such as (3) are negated is more interesting. In Akan, negation is realised as a homorganic nasal that is prefixed to a verb and is orthographically represented as /n/ or /m/. When (3a) and (3b) are negated the result is as in (18) and (19): (18)
?Na anka -n-kyere ade w kurom. Then anka s/he-NEG-teach thing be town. ‘It is not the case that she used to teach in town.’
(19)
?Kan no anka Oynyankop n m-men fam. Beginning DEF anka God NEG-close ground. ‘It is not the case that in the past God used to be close to the earth.’
c
c
c
As statements, (18) and (19) are grammatically odd. They are only acceptable as negative questions, which give a specific interpretation of S’s thought. In this case, the last syllable is produced on a relatively low pitch; the utterances will translate as (20) and (21). The declarative form represents the speaker’s belief (propositional attitude), while the negative polarity of the proposition expressed by (20) and (21) represents a thought attributed to someone other than the speaker. (20) (21)
She used to teach in town, didn’t she? God used to be close to the earth, didn’t He?
In (4), like in most of the cases in (1) and (2), an attempt to negate the anka-clauses result in pragmatically odd statements. The pragmatically (and sometimes grammatically) infelicitous use of negative with anka is the result of the inherent negation in the semantics of anka. Anka is used in representing a state of affairs which is not real and it is implausible to negate a non-existing state of affairs.
6. Summary In this paper, I have considered the various uses of Akan anka, (a) as a politeness marker, (b) in communicating unfulfilled desires and intentions, (c) in communicating the existence of a state of affairs which no longer exists, and (d) in communicating acts which were thought to have been very close to happening but did not actually happen. I have claimed that anka is not ambiguous between these different uses. Rather, it is vague, so that topdown extralinguistic inference is crucial in arriving at the intended interpretation. To consider the encoded meaning of anka as vague rather than ambiguous is a more economical and psychologically plausible way of accounting for this marker, considering the fact that in all its uses, it encodes the assumption that the proposition expressed by the
N.A.A. Amfo / Journal of Pragmatics 37 (2005) 997–1013
1013
utterance in which it occurs, is not factual. Because anka encodes a higher-level explicature, which conveys that (for speaker S, proposition P) P, in S’s opinion, is not factual, I call it a modal marker.
Acknowledgements I am grateful to Thorstein Fretheim and two anonymous reviewers for their careful reading and critical comments, which greatly improved the quality of this paper. Any remaining shortcomings are mine. An earlier draft of this paper was presented at the Linguistics Association of Great Britain (LAGB) Spring 2000 meeting. I am grateful to that audience as well.
References Amfo, Nana Aba Appiah, 2001. A relevance-theoretic study of some pragmatic markers in Akan. Working Papers in Linguistics, vol. 36. NTNU, Trondheim. Blakemore, Diane, 1987. Semantic Constraints on Relevance. Blackwell, Oxford. Blakemore, Diane, 2000. Indicators and procedures: nevertheless and but. Journal of Linguistics 36, 463–486. Blass, Regina, 1990. Relevance relations in discourse: a study with special reference to Sissala. Cambridge University Press, Cambridge. Christaller, J.G., 1933. A Dictionary of the Asante and Fante Language Called Tshi (Twi), 2nd edition Basel Evangelical Missionary Society, Basel (1st edition, 1881). Carston, Robyn, 2002. Thoughts and Utterances: The Pragmatics of Explicit Communication. Blackwell, Oxford. Clark, William, 1991. Relevance theory and the semantics of non-declaratives. Ph.D. thesis, University of London. Duranti, Alessandro, Goodwin, Charles (Eds.), 1992. Rethinking Context: Language as An Interactive Phenomenon. Cambridge University Press, Cambridge. Goffman, Erving, 1974. Frame Analysis: An Essay on the Organization of Experience. Harper and Row, New York. Grice, H. Paul, 1989. Studies in the Way of Words. Harvard University Press, Cambridge, MA. Obeng, Samuel G., 1999. Requests in Akan Discourse, 41. Anthropological Linguistics, 2. Sperber, Dan, Wilson, Deirdre, 1995. Relevance: Communication and Cognition. 2nd edition. With a Postface. Oxford, Blackwell. (1986). Wilson, Deirdre, Sperber, Dan, 2004. Relevance theory. In: Horn, Laurence R., Ward, Gregory (Eds.), The Handbook of Pragmatics. Blackwell, Oxford & Boston.
Nana Aba Appiah Amfo is a lecturer at the Department of Linguistics, University of Ghana and a PhD candidate at the Norwegian University of Science and Technology (NTNU). Her current research interests include the use and relevance pragmatic markers in communication, lexical pragmatics and the general use of language in society.