Emphasis, certainty, and interdiction: Adverbials in Gɩsɩɖa Anii

. This article presents novel fieldwork data describing the adverbials shɩ , cɔɔ , and caa in Anii, an understudied Ghana-Togo-Mountain language spoken in Togo and Benin. Shɩ and cɔɔ emphasize and mark the speaker’s certainty of aspectual reference, respectively, while caa serves to mark the speaker’s acknowledgement that the action or state denoted by the predicate is contrary to the listener’s wants.

1. Introduction.This paper, based on the undergraduate thesis "Emphasis, Certianty, and Interdiction: Particles in Gɩsɩɖa Anii" (Mariani 2023), presents novel fieldwork and the first analysis of the adverbials shɩ, cɔɔ, and caa in Anii, an understudied Ghana-Togo-Mountian language spoken in Togo and Benin.Shɩ is a marker of aspectual emphasis, and emphasizes the aspectual reference of the clause.Cɔɔ is a marker of aspectual certainty, and marks that the speaker is certain of the aspectual reference of the clause.Caa is a marker of interdiction and mockery, and indicates that the speaker acknowledges that the action or state denoted by the predicate is contrary to the listener's wants.Syntactically, I posit that these adverbials merge in the specifier position of the functional head they semantically relate to, following Cinque (1999).Shɩ and cɔɔ enter into a spec/head relation with AspP, and caa enters into a spec/head relationship with ModP.This analysis accounts for the semantic meaning of each adverbial, as well as accurately predicting the observed word order of the markers.

Background.
Anii is an understudied Ghana-Togo-Mountain language spoken in Togo and Benin-the data used for this project is of the Gɩsɩɖa dialect, which is spoken in Bassila, Benin (Figure 1).
Figure 1.Location of Bassila on a map of Africa and Benin (Eberhard, 2022), Local map credit: Martin Zaske 2.1.ASPECT, AKTIONSART, AND ANII.Anii has a somewhat complex system of marking temporal and aspectual reference, which involves the interaction between the Aktionsart of the predicate and overt morphemes marking imperfective (tə or na1 ) and perfect (cee) aspects (Morton 2014).
The interpretation of the aspectual reference in an Anii clause is not solely based on the aspect marker the clause may contain.The predicate Aktionsart, specifically the stative/eventive distinction, further restricts the potential aspectual interpretations of an utterance.For an unmarked clause, eventive predicates are only compatible with a perfective aspectual reference, and stative predicates are only compatible with an imperfective episodic aspectual reference.
Additionally, predicates with an overt aspect marker are still sensitive to Aktionsarten; Stative predicates marked for imperfective are restricted to the habitual, and eventive predicates marked for the perfect are restricted to the non-continuative perfect.In these cases, I treat the aspectual head to be a composition of both predicate Aktionsart and the overt aspect marker.These restrictions are summarized in Note that for this paper, I will be focusing on the perfective and imperfective aspects for brevity.See Mariani (2023) for discussion of the adverbials and the perfect aspect.Note as well that I am following Reichenbach (1947: Ch. 57), Klein (1994), and Dowty (1979) in defining aspectual reference.
2.2.CINQUE'S HIERARCHY.Cinque (1999) proposes a universal order of clausal functional heads, which is rigidly fixed by Universal Grammar.He further claims that Adverbial Phrases (AdvPs) enter into a spec/head relation with the head they are semantically related to.The order of the clausal heads therefore determines the order of the associated AdvPs.Cinque arrives at this conclusion and derives the hierarchy through cross-linguistic comparison.He first demonstrates that the functional heads have a universal ordering, based on data from languages which realize those heads through overt suffixes and particles.He then does the same for the relative order of AdvPs.Finally, he demonstrates that the clausal heads and AdvPs share a universal order, and claims that the AdvP enters into a spec/head relationship with the functional head that it is semantically related to.
Notably, Cinque's hierarchy was not developed with any data from Anii, however this article will demonstrate that it still forms accurate predictions about the order of clausal elements in Anii.Specifically, the adverbials shɩ and cɔɔ enter into a spec/head relationship with AspP, and caa enters into a spec/head relationship with ModP, with both positions accurately placing the adverbials after tense and prior to aspect.

Data.
The data presented in this paper was collected from April 2022 to February 2023 from five native speakers of Anii, and the language of elicitation was French.Elicitation sessions took place over Zoom video calls, and elicitation notes were shared via screen-sharing so that all participants could verify transcriptions.For the sake of clarity regarding phonological processes, data is presented in the orthography.
3.1.SHɩ: ASPECTUAL EMPHASIS.Shɩ, pronounced [ʃɪ ] is an adverbial used to give emphasis to the aspectual reference of the marked clause.Linearly, shɩ appears after tense marking (i.e., the future marker tə) and before any overt aspect marking.The adverbial does not act on temporal reference, and functions the same with or without the presence of an overt aspect marker.The following data demonstrate the usage of shɩ in various aspectual contexts, showing its apparent function as an aspectual emphasizer.
Consider the unmarked eventive predicate (1) below, which receives a perfective interpretation.In this case, the use of shɩ is interpreted as an emphasis on the completed nature of the event.
( In the above, the speaker chooses to emphasize the completion of sweeping because they are annoyed with the request, and perhaps to indicate that the listener should stop asking.This usage has a similar sense to the use of did as an emphatic in English.If, however, emphasizing the completed nature of an event is judged as unnecessary or excessive, shɩ is infelicitous.This is evident from the minimal trio in (2), where shɩ is only felicitous when the context licenses this emphasis.
( In (2), the use of shɩ is only felicitous if there is some reason to emphasize that mangoes were eaten.In (2a), the speaker is emphasizing that they are done eating mangoes, because they ate a lot of them, and are no longer hungry.In (2b), the speaker is emphasizing that they are done eating mangoes because they are sick, and do not wish to have any more.
With an overt aspect marking, shɩ still emphasizes the aspectual reference of the clause.Consider the imperfective-marked (3), in which shɩ serves to emphasize the habitual nature of the event -the fact that the speaker is habitually at home.While shɩ acts on aspectual reference, it does not act on temporal reference.For example, compare (4a) and (4b).In this instance, the emphasis imparted by shɩ is on the unbound nature of the speaker's being at home.The sense of uncertainty then follows from the difficulty of predicting the future, and not directly from the emphasis of shɩ.
( In summary, shɩ draws attention to and emphasizes the aspectual reference of the clause.It is only felicitous when there is some reason to emphasize the completion or unbounded nature of the event, and is not used to emphasize the temporal reference of the clause.
3.2.Cɔɔ: ASPECTUAL CERTAINTY.Cɔɔ, pronounced [tʃɔ ɔ] , is used to indicate that the speaker is certain of the aspectual reference.As with shɩ, cɔɔ appears before an overt aspectual marker or the verb, and is grammatical with all aspect markers.The following data demonstrates the usage of cɔɔ in various aspectual contexts, showing its apparent function as an aspectual certainty marker.
In the perfective, cɔɔ marks the speaker's certainty that the action denoted by the predicate has been completed.Consider (5), where the speaker is explaining that they have finally swept the courtyard.In elicitation, this interpretation was explained as the speaker pre-empting the listener's probable disbelief that they had swept the courtyard after a period of not being able to.This surfaces as a finally interpretation in translation.
With an overt marker, cɔɔ still acts on aspectual reference.Note the interpretation of 'definitely' in (6), signaling that the speaker is sure that the event in question is habitual.As with shɩ, cɔɔ does not interact with temporal reference.In (7), the speaker uses cɔɔ to mark that they are sure that they will be home for the entire time being discussed, which is in line with the imperfective habitual aspect.( 7 Above, I have demonstrated that shɩ and cɔɔ are adverbials that act on aspectual reference, and that they appear linearly between tense and aspect.Both adverbials have the same function regardless of the presence of an overt aspect marker.Below, I detail a third adverbial, caa, which does not act on aspectual reference.

CAA: INTERDICTION AND MOCKERY. Caa, pronounced [tʃáà]
, marks that the speaker acknowledges that the action/state denoted by the predicate is contrary to the listener's wants.While caa appears in the same linear position as shɩ and cɔɔ, it does not interact with aspectual reference.When using caa, there is often a strong connotation of mockery.Because of its deontic flavor (i.e., something that the speaker should not or cannot do), I posit that caa is an adverbial of modal necessity.The following data demonstrates the various usages of caa and its apparent function and meaning.
The most salient usage of caa is to indicate the willful disregard of a prohibition or interdiction given by the listener (9).If the listener is not opposed to the action/state denoted by the predicate, caa is infelicitous (9c).( 9 3.4.SUMMARY OF THE ADVERBIALS.To summarize the three adverbials, shɩ emphasizes the aspectual reference of the clause, cɔɔ marks the speaker's certainty of the aspectual reference, and caa marks the speaker's acknowledgement that the action or state denoted by the predicate is contrary to the listener's wants.All three adverbials appear in the same linear position after tense and before aspect.Shɩ and cɔɔ are aspectual adverbials, and act on the aspectual reference of the marked clause.Caa is a modal adverbial because of its deontic connotations, and does not act on aspectual reference. The above data establishes the meanings of the adverbials and their distributions in Anii.Next, I will detail how the adverbials fit into the clausal structure of Anii, and show that their positions and functions can be accurately predicted within the framework of Cinque's (1999) hierarchy.

Syntactic analysis.
With the meaning of shɩ, cɔɔ, and caa established, their structural position within the clause can be examined.The following section will present Cinque's (1999) hierarchy, and show that this framework accurately predicts both linear order and the semantic meaning of the adverbials.First, I will systemically narrow the hierarchy to only those heads which are compatible with the adverbials on syntactic grounds, and then I will use semantic meaning to determine which head each adverbial merges with.Finally, I will show that the proposed analysis correctly accounts for the observed data.4.1.CINQUE'S HIERARCHY.Cinque (1999) proposes a universal order of functional projections, and claims that adverbials enter into a spec/head relation with the functional head that they are semantically related to.The hierarchy is well supported with empirical data from cross-linguistic comparison.The full hierarchy is as follows: (13) Cinque's (1999) hierarchy: For the purposes of this analysis, the hierarchy can be narrowed in several steps.First, heads after tense (and the irrealis mood, see Morton and Blanchette (Under revision)) may be eliminated: (14) Cinque's hierarchy (narrowed): Then, heads that are irrelevant (i.e., not present as either heads or adverbials in Anii) may be eliminated and redundant (i.e., no distinction exists in Anii between adjacent heads) heads may be combined, giving the range of heads that may potentially host shɩ, cɔɔ, and caa: (15) The range of functional heads capable of accommodating the adverbials shɩ, cɔɔ, and caa: Finally, the volitional modal may be eliminated as it would cause an ordering issue with the imperfective and perfective aspect markers, and adjacent aspectual heads may be combined: (16) Potential functional heads for shɩ, cɔɔ, and caa: Mod necessity < Mod possibility < Asp 4.2.MAPPING THE ADVERBIALS TO FUNCTIONAL HEADS.Now that the hierarchy has been narrowed down to heads that are compatible with shɩ, cɔɔ, and caa, the mapping of each adverbial to the correct functional head is semantic in nature.The head that each adverbial is associated with will select for the adverbial as its specifier.
In the case of shɩ and cɔɔ, their close relation to aspectual reference motivates their position as the specifier of aspect.There is no motivation to merge them as modals, as they have no deontic, dynamic, or bouletic interpretation.Caa, on the other hand, does not interact with aspectual reference, making a spec/head relation to aspect unlikely.Instead, I posit that caa merges under modal necessity, due to its deontic interpretation that it is something the speaker should or cannot do from the viewpoint of the listener.
To summarize, the only heads that are compatible with shɩ, cɔɔ, and caa are modal necessity, modal possibility, and aspect.Due to shɩ and cɔɔ having a close semantic relation with aspectual reference, I hold that these merge as the specifier of AspP.As caa has a strong deontic interpretation, I argue that it merges as the specifier of modal necessity.Within the framework of Cinque's hierarchy, these positions would accurately account for the positioning of the adverbials after tense and prior to aspect.4.3.SYNTACTIC REPRESENTATION.The above sections demonstrate that shɩ and cɔɔ are structurally located as the specifier of AspP, and that caa is in the specifier position of ModP.This section will model this representation and show that the proposed structure accurately predicts the patterns observed in the data.Caa, on the other hand, merges into the specifier position of ModP.Because this position is between tense and aspect, the linear order of the clause is maintained, while caa is in a position related to its deontic interpretation (18).This structure models both the order and the semantic meaning of the adverbials.In addition, it is extensible -if other adverbials are discovered that act on clausal heads, their order may be able to be modeled with a similar Spec/Head relation in the order predicted by Cinque's hierarchy.4.3.2.PREDICTIONS OF LINEAR ORDER.The structure presented above accurately generates the attested linear order of Anii clauses containing the adverbials.This is shown in (19-21) below for the aspectual adverbials.Notice that the adverbials are predicted to appear in the correct linear order for non-future sentences without (19) and with (20) overt aspect marking, and for future sentences (21).Note also that I am treating the "future construction" n tə ́as a single unit for the purposes of this analysis.
The proposed structure also accurately predicts the linear order of clauses containing caa while accounting for its deontic meaning.As with the aspectual adverbials, this is true for nonfuture sentences without (22) and with (23) overt aspect marking, and for future sentences (24).The above examples illustrate that the proposed structures can accurately account for the word order of the clause.In addition, the semantic meanings of shɩ, cɔɔ, and caa are accounted for by the Spec/Head relationship to the associated functional head.Combining the two structural "maps" (17) and ( 18

5.
Conclusions.Within this paper, I have presented the adverbials shɩ, cɔɔ, and caa and detailed their meaning, usages, and syntax.This paper has presented novel data on previously unstudied markers, as well as providing the first analysis of adverbial syntax in Anii.This analysis concludes that shɩ is a marker of aspectual emphasis, cɔɔ is a marker of aspectual certainty, and caa is a marker of interdiction and mockery.Syntactically, shɩ and cɔɔ enter into a spec/head relation with Asp, while caa enters into a relationship with Mod necessity .This analysis accounts for the semantic meaning of the adverbials, and also accurately predicts the word order they surface in.The results of this paper are summarized in ( 26) and ( 27

IMPLICATIONS AND FURTHER WORK
. This paper represents the first inquiry into the adverbials shɩ, cɔɔ, and caa, and contributes a substantial amount of data to the literature on Anii.In addition, it presents an analysis of adverbials in Anii, which has never been examined in detail prior to this work.The data and analysis presented herein will provide a starting point for the analysis of similar markers in Anii and in other languages.Additionally, since Anii was not one of the languages used by Cinque (1999) to create the hierarchy, this analysis extends Cinque's theory to a new language.This provides additional empirical support for placing adverbials in the specifier of the head they modify.
In terms of future inquiry into the adverbials shɩ, cɔɔ, and caa, a formal semantic analysis of their emphatic and pragmatic nature is a potential avenue for a better understanding of Anii and aspectual reference.In addition, a wider scoping analysis of adverbials in Anii may allow for a more robust comparison of relative adverbial positioning and point more directly to taking either a scopal or cartographic approach to their placement.In addition, there is a possibility of similar makers and emphatic particles in other languages, which may provide more insight as to how the languages in this region and family mark tense, aspect, and adverbials, and how these phenomena may be modeled.All of these are open to further inquiry, and with time more insights will be gained about Anii, adverbials, and syntax as a whole.
(3) (Imperfective habitual) a. Context: You are waiting for a visitor, but they haven't told you what day they'll arrive.So, you are at home each day.
(5) (Perfective) a. Context: Someone asks you what you did the day before, and you swept the courtyard.You reply: n 1SG shɛr sweep gaka courtyard 'I swept the courtyard.'b.Context: Someone asks you what you did the day before, and you swept the courtyard after meaning to do so for a while.You reply: swept the courtyard.' (6) (Imperfective habitual) a. Context: Someone asks where you normally spend your time.You are usually at home, so you respond: at home.' b.Context: Someone asks where you normally spend your time.You are usually at home, so you respond: n 1SG cɔɔ CERT tɩ IMPF ɖa be.there afal house 'I definitely stay at home.' 4.3.1.THE STRUCTURAL POSITIONS OF THE ADVERBIALS.Shɩ and cɔɔ, the aspectual adverbials, merge into the specifier position of AspP (17).This position accurately predicts the positioning of the adverbials below tense and in a structural position related to aspect.(17)Position of AdvP within AspP: ) yields the final structural positions below.(25)The Relative Positions of shɩ, cɔɔ, and caa: ) below: (26) Definitions of shɩ, cɔɔ, and caa: shɩ: Emphasizes the aspectual reference of the marked clause.cɔɔ: Marks certainty of the aspectual reference of the marked clause.caa: Marks the speaker's acknowledgment that the action or state denoted by the predicate is contrary to the listener's wants.(27) The Relative Positions of shɩ, cɔɔ, and caa:

Table 1 .
Aspectual references of Anii clauses With shɩ, the statement is interpreted with a sense of uncertainty due to the emphasis on the hard to predict nature of the future imperfective.With cɔɔ, however, the predicate is clearly interpreted with a sense of certainty, separating it from emphasis.You see a friend at the market and they ask what you will be doing that night, since they tried to visit you and couldn't find you.You will be at home and are sure of it, so you respond: 'I will probably be at home then.' 3.2.1.COMPARISON WITH SHɩ.Because both appear in the same linear position and both act on aspectual reference, shɩ and cɔɔ are difficult to tell apart, even for our consultants.The difference between the emphasis of shɩ and the certainty imparted by cɔɔ is more clearly visible in the future.Compare (8a) and (8b).
) a. Context: Someone calls you and asks where you normally spend your free time.Your sister is not allowed to touch the TV because she is young and may break it.You go out, and when you return she is watching TV and does not see you.You stand there watching her until she hears a noise and rushes to turn it off and attempts to act nonchalant.You tell her that you've been there for a while:Because caa is an explicit acknowledgement that the speaker knows that the listener's wants are violated, it often carries a strong connotation of mockery.Note as well that caa is incompatible with a sense of apology or sincerity (11b).(11)a.Context: A brother and sister are playing, and the sister slaps the brother from behind.(She is proud of this.)The brother exclaims 'What was that?', and the sisterThe sense of mockery is mostly lost in the future, as the speaker is not simultaneously violating the listener's wants and pointing it out to the listener (12) (12) Context: Someone asks you to go out but you are sick and think it might be better for you to stay at home, so you reply: