Dutch individuals with Broca’s aphasia and their comprehension of epistemic (un)certainty with or without visual evidence

(1)

Graduate School for Humanities MA Program in General Linguistics

Dutch individuals with Broca’s aphasia and their comprehension of

epistemic (un)certainty with or without visual evidence

Thesis submitted by Jojanneke Drapers (student number: 6036902) Supervisors: mw. dr. Tuba Yarbay Duman mw. prof. dr. Jeannette Schaeffer

(2)

2

Abstract

Background: Several studies have shown that individuals with Broca’s aphasia have difficulties with the production and comprehension of verbs that refer to the past. Verbs that refer to the past are not only verbs in past tense, like walked, but also verbs referring to the past by participles, like in he has thrown a ball. Referring to the past through a verb form expresses epistemic certainty: certainty about the happening of the event. However, as the event being referred to happened in the past, often there is no evidence available anymore to support this certainty. Yarbay Duman and Bastiaanse (2009) hypothesized that possibly, individuals with Broca’s aphasia have difficulty interpreting epistemic certainty when there is no visual evidence available. Aims: The aim of this study is to test whether participants with Broca’s aphasia are able to comprehend the Dutch modal auxiliaries kunnen (which expresses epistemic possibility) and moeten (epistemic certainty) with or without visual evidence. Another aim is to find out whether these participants are able to differentiate between varying levels of epistemic certainty when no words are used to express this. Methods and Procedure: A test was designed, based on the study by Ozturk and Papafragou (2015) in which 10 participants with Broca’s aphasia and an age-matched group of 16 healthy adults were presented with short stories in which was hinted on the location of an animal. There were four story variations that provided either indirect positive visual evidence (the participant could infer that the animal should be in a box) or no visual evidence (with two closed boxes, no conclusions could be drawn about the location of the animal). The participants were then offered statements about the location of this animal. This statement always included kunnen (the kunnen condition) or

moeten (the moeten condition). Also a condition was designed to test the capabilities to

differentiate different levels of epistemic certainty without words that express epistemic modality; in this condition, participants were asked to point to one picture out of three in which they were most certain that the animal is in a certain box. Results and Outcomes: Individuals with Broca’s aphasia performed significantly worse than healthy adults in situations with no visual evidence when the offered statement included the Dutch modal verb moeten, which expresses epistemic certainty. No difficulties were found for situations where visual evidence was available; also no difficulties were found for items in the kunnen conditions. Furthermore, the participants did not appear to have difficulty differentiating between varying levels of epistemic certainty without words that express epistemic modality. Conclusions: The idea that individuals with Broca’s aphasia have trouble interpreting epistemic certainty when no visual evidence is available, is supported by this study. Also did this study show that individuals with Broca’s aphasia are not significantly worse in interpreting varying levels of certainty from situations in which no words are used to express epistemic certainty. However, different response patterns were found, which suggests that it could be a cognitive deficit. This last finding supports data from Yarbay Duman (in press), that suggested that participants with brain injury had difficulty with this epistemic certainty is not specific to language. More research is needed on this. Either way, epistemic certainty appears to be problematic in individuals with Broca’s aphasia and could be the foundation of the problems that are repeatedly found with verbs that refer to the past.

(3)

3

1 Introduction

Aphasia is a brain disorder resulting in difficulty producing and/or comprehending language, mostly caused by acute brain injury such as a cerebrovascular accident (CVA) (Faroqi-Shah and Thompson, 2007). Broca’s aphasia is an aphasic syndrome which most strikingly affects production: omission of function words, word finding problems are most salient. Furthermore, individuals with Broca’s aphasia also often show problems with verb inflection (Friedmann and Grodzinsky, 1997; Bastiaanse and Thompson, 2005; Clahsen and Ali, 2009; Faroqi-Shah and Dickey, 2009; amongst others) and sentence comprehension (Caramazza and Zurif, 1976; Schwartz, Saffran and Marin, 1980; Saffran, Schwartz and Marin, 1980; amongst others). These problems have long been explained by a deficiency in morphology (Faroqi-Shah and Dickey, 2009; Clahsen and Ali, 2009) or difficulties with complex syntactic structures in which these verbs are often found (Caramazza and Zurif, 1976; Saffran, Schwartz and Marin, 1980; Friedmann and Grodzinsky, 1997; Bastiaanse and Thompson, 2005).

Recently, several studies have shown that individuals with Broca’s aphasia mostly have problems with the production of verbs that refer to the past (Bastiaanse, 2008; Yarbay Duman and Bastiaanse, 2009; Bastiaanse et al., 2011); not only for past tense inflection like walked, but also participles such as ‘he has thrown the ball’, where no past tense is involved and the reference to the past is expressed by a participle. Bastiaanse (2008) and Bastiaanse et al. (2011) gave a morphological explanation for these data: there could be an underlying deficit in reference to the past through grammatical morphology.

However, Yarbay Duman and Bastiaanse (2009) considered that the problem might be semantic rather than morphosyntactic. According to the authors, referring to the past through a verb form inherently expresses epistemic certainty: by referring to a past event through a verb form, the speaker inherently states that the said event indeed has happened. They hypothesized that individuals with Broca’s aphasia could have difficulty with the assignation of epistemic certainty (the degree of certainty a speaker has about the likeliness of an event to happen) to morphosyntactic structures. When they are uncertain as to the happening of the event, they use visual evidence to determine whether the event has happened or not (see also Yarbay Duman and Topbas, 2016).

When a speaker refers to an event through a past verb form, i.e. to an event that has taken place in the past, there is often no (visual) evidence available anymore. For example, when one

(5)

5

says that ‘John has eaten the cookies’, the cookies are gone; or in the case of ‘I have been dancing yesterday’, there is generally no evidence left of this. A verb form that refers to the past, is a cue that an event has been completed, and simultaneously implies certainty that the event has happened (i.e. provides epistemic certainty).

In some other situations however, there can be evidence left, from which the happening of the event can be inferred; for example, when the streets are wet, one can say ‘it has been raining’. These wet streets form indirect visual evidence for the proposition that it was raining earlier, as the event itself is not being witnessed, but merely the results of it. The speaker connects these results to the (inferred) event.

When talking about an event in the future, one inherently expresses uncertainty; as the event has not happened yet, one can never be absolutely certain that the event indeed will happen (Nuyts, 2011: “Situating something in the future unavoidably involves an element of uncertainty”; p.173). Meanwhile, there is often relatively more visual information or evidence available for the potential happening of the event. For example, when one says ‘It is going to rain’, one could rely on black clouds in the sky.

When the event is happening in the present, the speaker is often looking directly at the event as it is happening; let us take the rain as an example again. When someone looks outside and sees the rain coming from the sky, he can say ‘it is raining’. It is generally easy to find evidence for an event that is happening in the present (Yarbay Duman and Topbas, 2016). Furthermore, a verb that refers to the present, does not have a completion cue (as the event is still happening), so also does not imply that the event has happened. Therefore, epistemic certainty is irrelevant for events that happen in the present.

Dutch modal verbs can explicitly express epistemic possibility (kunnen) or epistemic certainty (moeten). If it is the case that individuals with Broca’s aphasia have difficulty with epistemic certainty as expressed in verbs that refer to the past, then verbs expressing epistemic certainty could also be expected to be problematic. The aim of this thesis is therefore to test whether this is the case.

This leads to the following research question:

Do individuals with Broca’s aphasia have trouble comprehending the modal verb moeten when visual evidence is or is not available?

Modal verbs have not been tested extensively for individuals with Broca’s aphasia before. Yarbay Duman (in press) tested three participants with brain injury and showed that they had

(6)

6

difficulty with epistemic inferencing: they appeared to be uncertain about an event for which there was no direct perceptual evidence and to need visual evidence to be able to be certain that the event has happened.

The hypothesis by Yarbay Duman and Bastiaanse (2009) states that individuals with Broca’s aphasia have difficulty with the interpretation of epistemic certainty when no visual evidence is available. This thesis tests that hypothesis by the means of Dutch modal verbs kunnen (epistemic possibility) and moeten (epistemic certainty), in a sentence comprehension task with a participant group of Dutch monolinguals with Broca’s aphasia and a group of Dutch healthy adults.

No apparent difficulty is predicted to be found with moeten when visual evidence is available; also no apparent difficulty is predicted to be found with the modal verb kunnen, whether visual evidence is available or not. Also, not only words that express epistemic certainty are assumed to be problematic; individuals with Broca’s aphasia often have cognitive problems and Yarbay Duman (in press) found that individuals with brain injury also have difficulty with epistemic inferencing without linguistic means. Therefore, it is hypothesized that they will not have problems with epistemic certainty when no words that express that notion are involved.

In the next chapter, modal verbs will be defined and discussed more extensively to draw a complete picture of the word class. Modality in general, and epistemic modality will be explained, together with relevant notions for this study. Also aphasia, and more specifically Broca’s aphasia, together with relevant studies for this thesis will be mentioned. These relevant studies will mostly concern typically developing (TD) children, as there has not been extensive research on modal verbs in Broca’s aphasia before.

2 Theoretical background

2.1 Epistemic modality

Modality is the qualification of a state of affairs as expressed in language (Nuyts, 2001) or, stated differently, modality is concerned with the opinion and attitude of the speaker (Lyons, 1977; as mentioned in Palmer, 2014). The notion of modality in itself is not a linguistic one,

(7)

7

but merely a cognitive one; evaluating the likeliness of a state of affairs appeals to cognitive capabilities (Nuyts, 2001). However, the expression of modality in language is a linguistic topic, and also the topic of this thesis; more specifically, this thesis treats the comprehension of the expression of epistemic modality by individuals with Broca’s aphasia.

Nuyts (2001) mentions three categories of modality: (1) epistemic modality; (2) deontic modality and (3) dynamic modality. First, epistemic modality will be explained; after that, deontic and dynamic modality will be explained briefly to make clear the differences and clarify the notion of epistemic modality.

Epistemic modality is the speaker’s evaluation of the likelihood of a state of affairs to happen (or to have happened), based on ‘what is known’ (about the real world or any other world). For example, the sentence ‘maybe they have run out of gas’: when it is known ‘they’ were coming by car, and it is also known that they have not yet arrived, it is a possibility that they have run out of gas. However, there is no certainty that this is the case; there is not enough evidence to say, for example, ‘certainly, they have run out of gas’. In this case, more or clearer information is known by the speaker than when he uttered the previous sentence expressing epistemic possibility; for example, the driver of the car has called earlier to tell they have almost ran out of gas. This shows that (epistemic) modal expressions often form a scale that goes from weak to strong; from possibility to probability to certainty.

This also counts for the other two kinds of modality: Deontic modality and dynamic modality. Deontic modality is an evaluation of the moral acceptability or desirability of a state of affairs (Nuyts, 2001); i.e. an evaluation of what is acceptable or desirable. An example of deontic modality is ‘you are allowed to leave’. This example expresses deontic possibility; the other person is given a choice. To express deontic necessity, one could say ‘you need to leave’, which is a stronger expression.

Finally, dynamic modality expresses an ‘inherent capacity of an agent participant in a state of affairs’ or an ‘inherent potential of a situation to allow an agent participant to do something’ (Nuyts, 2001). Important about this kind of modality is that the state of affairs is always inherent regarding a subject or situation; which is not the case in epistemic and deontic modality. Consider for example: ‘to get to the city center, you need to take bus 34’. This sentence expresses a necessity to take bus 34 to achieve the goal; this necessity is inherent to the goal. One could however also express dynamic possibility by saying ‘to get to the city center, it is

possible to take bus 34’. In this case, multiple buses go to the city center; bus 34 is just one of

(8)

8

There are three ways of expressing epistemic modality: (1) by using adjectives or adverbs; (2) by using mental state predicates; (3) by using modal auxiliaries (Nuyts, 2001). See sentences (1)-(3) for examples that subsequently express epistemic modality through an adjective, an adverb, and a mental state predicate. These examples are all borrowed from Nuyts (2001).

(1) It is probable that they have run out of fuel. (2) Probably they have run out of fuel.

(3) I think that they have run out of fuel.

The current study focuses on the third way mentioned by Nuyts (2001): modal verbs (i.e. modal auxiliaries). Modal verbs are a specific kind of auxiliary, and an auxiliary is a specific kind of copular verb.

Verbs  copular verbs  auxiliaries  modal auxiliaries

Copular verbs form another class of words, a closed class of words (i.e. function words). Copular verbs link a subject to something else: another verb, a state, or a characteristic. In English, to be is the most frequently used copular verb. See sentence (4) for an example; here, the copular verb is used to connect the subject ‘he’ to the state ‘very sad’.

Auxiliaries are a subset of copular verbs, which must, unlike regular copular verbs, always link the subject with another verb. One of the functions of auxiliaries is to make a sentence refer to the past; see also sentence (5).

(4) He is very sad that his girlfriend left him. (5) I have called my mother yesterday.

Modal auxiliaries are a specific kind of auxiliary, which express either epistemic, deontic or dynamic modality. In many languages, including English and Dutch, one set of modal verbs is used for all three kinds of modality, while epistemic modal adverbs and mental state predicates do not have multiple meanings; they just express epistemic modality.

In the current task, Dutch modal verbs will be used to express epistemic modality: Dutch

kunnen (‘can’/‘might’) behaves like maybe, i.e. expresses epistemic possibility; while Dutch moeten (‘must’/’have to’) behaves like certainly, i.e. expresses epistemic certainty. In the

(9)

9 2.1.2 Modal verbs in Dutch

In this study, Dutch native speakers will be tested on their comprehension of modal verbs. Therefore, Dutch modal verbs will be used in the test items. In this section, properties and examples of epistemic modal auxiliaries in Dutch will be discussed.

Dutch has an SOV (subject – object – verb) word order and a Verb Second-rule, which means that the finite verb or auxiliary (so also modal auxiliaries) moves to the second position in the matrix sentence.

Unfortunately, there is very little literature specifically on Dutch modals. Nuyts (2001) did an extensive study on how epistemic modality behaves in English, German and Dutch, which was the only study that included Dutch modal verbs that was relevant for this study.

Luckily, Dutch modal verbs are very much comparable to English modals, as the languages are very closely related. In both languages, there is only one set of modal verbs that can be used to express all three kinds of modality (i.e. epistemic, deontic and dynamic). In English, this set of modal verbs includes can, may, might (epistemic possibility); must and have to (epistemic certainty). Dutch has an even smaller set of modal verbs: only kunnen; moeten and zullen (Nuyts, 2001).

Moeten is generally viewed as the Dutch counterpart of English must and kunnen as the

counterpart of may or can. Zullen expresses also a strong futurity value, like will (Cornillie, 2009; Nuyts, 2001; amongst others). Dutch modal verbs can also be combined to form ‘conditional’ forms (Mortelmans, Boye and Van der Auwera, 2009). See table 1 for an overview of the Dutch modal auxiliaries and their English counterparts.

Dutch English

Kunnen Can / may

Moeten Must / have to

Zullen Will

Zou kunnen Could Zou moeten Ought

Table 1: frequently used Dutch modal verbs and their English counterparts

(10)

10 (6) Hij kan wel eens ziek zijn geworden.

‘He may have become sick’

(7) Je portemonnee moet wel in de keukenla liggen. ‘Your wallet must be in the kitchen drawer’ (8) De bel gaat – dat zal Jan zijn.

‘The bell rings (lit. goes) – that will be Jan’

(9) Het zou kunnen dat ze geen benzine meer hebben. ‘It could be that they have run out of fuel’

(10) Het zou goed moeten smaken voor dat geld. ‘It ought to taste good for that money’

This set of modal verbs (Kunnen, moeten, zullen) can however also be used for expressing other concepts. Moeten is primarily deontic and also inferential, so has a clear evidential meaning (source of information; see also section 2.1.4); kunnen is primarily dynamic as it expresses ability; and about zullen there is some discussion. According to Coates (1983; Nuyts 2001) will can be used in two different ways: as a marker of future tense (‘prediction’ sense) and in a ‘predictability’ sense. Zullen behaves similar to English will. See for examples sentences (11) and (12):

(11) De bel gaat – dat zal Jan zijn.

“The bell rings (lit. goes) – that will be Jan”

(12) Ik zal je missen. “I will miss you”

In sentence (6), the speaker bases his proposition (the person opening the door being Jan) on what he knows, i.e. ‘what is known’ (Jan was the only person that was still to arrive, for example), and therefore clearly expresses epistemic modality. Example (7) however, seems to predict what it will be like for the speaker to see the other person/when the other person leaves. Referring to the future inherently expresses epistemic uncertainty; as one can never know with absolute certainty what will happen in the future. However, zullen expresses certainty; it conveys no doubt about the proposition. This example therefore does not express epistemic uncertainty by itself, but situates an event in the future; and doing this unavoidably carries epistemic uncertainty (Nuyts, 2001). As zullen is mostly used in this futurity meaning, this can be said to be the primary use of zullen.

(11)

11

Also the modal auxiliaries that will be used in the current task, kunnen and moeten, have multiple readings apart from their epistemic certainty reading: they can express also both deontic and dynamic modality. See examples (13), (14) and (15) for respectively an epistemic, deontic and dynamic reading of moeten.

(13) Hij moet zijn gevallen “He must have fallen.” (14) Je moet nu weg.

“You must leave now.”

(15) Om in het centrum te komen, moet je bus 34 nemen. “To get to the city center, you must take bus 34.”

In the examples above, moeten can be replaced by kunnen (‘might’), and this will in each sentence express the ‘weaker end’ of each kind of modality. In example (13), the speaker evaluates what he knows; he possibly sees a wound on the knee of the person he speaks about. Example (14) is an evaluation of what the speaker thinks is desirable: he wants the other person to leave. In sentence (15), taking bus 34 is inherently necessary to get to the city center.

These multiple meanings of (epistemic) modal verbs normally do not cause ambiguity; as in all words with multiple meanings, context generally makes clear which meaning is being targeted. If not, it is either not important which meaning is intended, or a miscommunication would occur, in which case the listener would ask for the necessary information.

For Dutch modal auxiliaries, the combination of a modal auxiliary and the verb zijn (to be) in an existential way often is an indication for an epistemic reading of the modal (Nuyts, 2001). However modal auxiliaries do generally not carry a primarily epistemic meaning, but a rather deontic or dynamic one, the situation and other components in a sentence make an epistemic reading more likely. In the current test, the context of the test will make an epistemic reading of the verbs unavoidable.

2.1.3 Semantic and pragmatic theory on (modal) scales

It has been mentioned and shown in previous sections that modal verbs kunnen and moeten form a scale that goes from weak to strong, where kunnen is the weaker end of the scale and

moeten is the stronger end. Scales like epistemic modal scales have two different possible

(12)

12

show the semantic interpretation, while adults more often show the pragmatic interpretation. In the current study. scales will not be tested, but both interpretations will be possible in some items in the task, as kunnen and moeten will be used in different situations. It is important to keep that in mind while analyzing the results. In this section, both the semantic and pragmatic interpretation of scales will be explained briefly.

According to the standard semantic analysis on these scales, a stronger statement entails a weaker one (Horn, 1972); this view is also mentioned by Noveck (2001), discussing the neo-Gricean view on scales like this. Or, as Papafragou and Musolino (2003) state: “some means

some and possibly all”; which is a more clear description of the semantics of scales. When

someone saw all my friends today, he could say he saw some of them and still be logically right; but when he saw some of them, it would not be right if he said he saw all.

Some  All

For modal verbs, this means that might/can (‘kunnen’) does not exclude must (‘moeten’); so when something must be the case, it would still be acceptable to say it might be the case; however, when something might be the case, it is not acceptable to say it must be the case. Ozturk and Papafragou (2015) showed that adults show this interpretation when not given the option of a pragmatic analysis (see section 2.2.1 for a description of their study). This semantic interpretation is the interpretation that will be viewed as default in the current test.

However, many adults (numbers varied in different studies; see also section 2.2.1) would interpret might as must not (Noveck, 2001; Noveck, Ho and Sera, 1996; Papafragou and Musolino, 2003). In other words, kunnen (‘can’/’might’) and moeten (‘must’) are mutually exclusive in this interpretation. This implication (might implies must not) is called a scalar

implicature; a weaker statement implies that a stronger one is not true.

Most of the studies on modal verbs, actually investigate scalar implicatures. In section 2.2.1, some studies will be discussed that show how children interpret modal verbs, and that they only learn to pick up on scalar implicatures at a later age. These studies also showed that adults tend to show a pragmatic reading of modal verbs. In this study, scalar implicatures are not the topic of investigation. It is however unavoidable to mention them, as adults, so probably also adults with Broca’s aphasia, tend to pick up on them when possible, and it is necessary to take this into account in this study. See also section 2.2.1 and the Discussion section.

(13)

13 2.1.4 Types of evidence and the use of modal verbs

It has been mentioned above that epistemic modality is the evaluation of the likeliness of a state of affairs based on ‘what is known’; and what is known can come from different sources of information, i.e. types of evidence. Some types of evidence are direct evidence (where something has been directly observed or experienced); inferential (where the happening of the event has been inferred by the speaker); and hearsay (when someone else has told the information to the speaker) (Whitt, 2009). Inferential and hearsay evidence could both be classified as indirect evidence: as it was not directly observed by the speaker.

This type of evidence that will be used in the current test is direct visual evidence and indirect visual evidence (i.e. inferential evidence). In the introduction, the ‘rain’ example has already been introduced and the idea that verbs referring to different moments in time express different amounts of epistemic certainty (including that epistemic modality is irrelevant in the case of present tense) has been explained.

As epistemic certainty is irrelevant in the case of events that are happening in the present (and for which visual evidence is available), one would normally not use a modal verb while referring to such an event. It would seem odd to say ‘it must be raining’ while the speaker is still looking at the rain coming from the sky. ‘What is known’ is what is being observed directly: that it is raining. A proposition like ‘it is raining’, without modality, is called a bare proposition. When indirect visual evidence is available, however, and an event has to be inferred to either have happened in the past, or to happen in the future, it is appropriate to use a modal verb in the sentence. This also means that when a modal verb is being used in a sentence, it generally expresses less certainty than when a bare proposition is being uttered.

This use of different kinds of source of information (evidence) is also concerned with another linguistic notion: evidentiality. Evidentiality is the expression of information source of the proposition (Aikhenvald, 2004). Epistemic modality is often mentioned in combination with evidentiality, because of the fact that they both use known information and/or the way of acquiring this information. However, they are two different concepts. Broadly speaking, evidentiality just states a source of information, without any judgment of the value of the information; while epistemic modality expresses amount of certainty and, as has been discussed above, truth value and/or commitment of the speaker. Clearly, the two notions are at least overlapping; as in the current thesis I will treat the modal auxiliary must as an epistemic modal auxiliary, it can be used (and is often used) as an evidential marker too, as a hearsay marker.

(14)

14

See sentence (16) for an example, given to highlight the difference between epistemic modality and evidentiality.

(16) That must be a good movie.

The epistemic modal interpretation of the verb must (or moeten in Dutch) was explained above. The evidential (hearsay) interpretation could be explained by the following situation: the speaker of sentence (16) is telling someone about a certain movie. He knows a friend of his has already seen that movie, and told him it was very good. He can then utter sentence (16), his source of information being what his friend told him (hearsay).

Because of their obvious closeness, evidentiality and epistemic modality have been combined in two ways in the literature: by inclusion, which means evidentiality is a subcategory of epistemic modality, or vice versa (Palmer, 1986; Plungian, 2001); and overlap, which means that part of the meaning of both notions is virtually the same (Van der Auwera and Plungian, 1998). In the current thesis, however, no evidentiality is being tested. Evidentiality has been mentioned briefly to explain the difference between the two concepts and clarify the notion of epistemic modality.

In short, direct visual evidence generally causes the speaker to utter a bare statement (without a modal verb): epistemic modality is irrelevant then. This is generally the case when an event is situated in the present. When indirect visual evidence can often be used to infer that the event has happened (or will happen). In these cases, a speaker often uses a modal verb. To situate an event in the future in Dutch however a modal verb can be used: zullen (as in English will is needed); these languages do not have a simple form to refer to an event in the future.

2.1.5 Summary of epistemic modality

Epistemic modality is the qualification of a state of affairs as expressed in language. It can be expressed by modal auxiliaries, amongst others. The two modal auxiliaries that in Dutch express epistemic possibility and epistemic certainty are kunnen (possibility) and moeten (certainty). Modal verbs in Dutch can be used to express all three types of modality: epistemic modality, but also deontic and dynamic modality.

Words that express (epistemic) modality form a scale that goes from weak (kunnen; possibility) to strong (moeten; certainty). Semantically, kunnen does not exclude moeten. However, many adults show a pragmatic interpretation of these kinds of scales; when

(15)

15

something can be the case, this implies that it must not be the case. This implication is called a

scalar implicature.

According to Yarbay Duman and Bastiaanse (2009), referring to the past through a verb form expresses epistemic certainty: certainty that the said event has happened. However, as the event has happened in the past, there is often no visual evidence available (anymore) to support this epistemic certainty. In other situations, there is (indirect) visual evidence available from which the happening of the event can be inferred and can support in the epistemic certainty.

When one situates an event in the future, this inherently expresses epistemic uncertainty or possibility, as one can never know for sure that this event will indeed happen; however, there is generally some indirect evidence available (e.g. black clouds in the sky, for the proposition ‘it is going to rain’). For events that are happening in the present, epistemic certainty is irrelevant, as there is often direct (visual) evidence available for the happening of the event. See table 1 for an overview of what kind of evidence is generally available when events are situated in past, present or future, and whether they express epistemic certainty.

Table 1: overview of availability of visual evidence and relevance and expression of epistemic certainty

of verbs.

Individuals with Broca’s aphasia often show a deficit in the production and comprehension of verbs that refer to the past (Bastiaanse, 2008; Yarbay Duman and Bastiaanse, 2009; Bastiaanse et al., 2011). Yarbay Duman and Bastiaanse (2009) hypothesized that epistemic certainty, i.e. the certainty that an event has happened, might be a problem when no visual evidence is available to support this certainty. In the following section, Broca’s aphasia will be discussed more extensively.

Event situated in the…

Past Present Future

Visual evidence available? None

(sometimes indirect)

direct indirect

(16)

16 2.2 Aphasia

The research question of the current study is whether individuals with Broca’s aphasia have trouble comprehending words that express epistemic modality (moeten, which is a Dutch modal auxiliary) when visual evidence is or is not available. In the previous section, epistemic modality has been explained; in the current section, Broca’s aphasia will be discussed more extensively to make clearer why this study focuses on epistemic modality in this type of aphasia.

Aphasia is a brain disorder resulting in difficulty producing and/or comprehending language, mostly caused by acute brain injury such as a cerebrovascular accident (CVA). Broca’s aphasia is an aphasic syndrome in which production of language is impaired most clearly. The most well-known symptoms are non-fluent production, and omission of function words and grammatical morphemes, also called ‘agrammatism’ (Goodglass, 1976); and difficulty with the comprehension and production of sentences with a non-standard word order (Caramazza and Zurif, 1976; Schwartz, Saffran and Marin, 1980; Saffran, Schwartz and Marin, 1980). Verb inflection is found to be a vulnerable category in Broca’s aphasia (Friedmann and Grodzinsky, 1997; Bastiaanse and Thompson, 2005; Clahsen and Ali, 2009; Faroqi-Shah and Dickey, 2009; amongst others).

In earlier studies, these difficulties with verb inflection were assumed to be a morphological or a syntactic problem. Friedmann and Grodzinsky (1997) argued that the tense node in the syntactic tree is unavailable in agrammatic speakers, and Faroqi-Shah and Dickey (2009) brought up a morphosemantic account which entails that the diacritic tense features are affected.

However, more recently, the idea has risen that the problem with verb production in aphasia could be semantic in nature. Several studies have shown that the problems of individuals with Broca’s aphasia mostly occur with verbs that refer to the past (Bastiaanse, 2008; Yarbay Duman and Bastiaanse, 2009; Bastiaanse et al., 2011); not only for past tense inflection, but also participles such as ‘he has thrown the ball’, where no past tense is involved; the reference to the past is expressed by a participle.

The studies mentioned above have shown that Dutch, Chinese, English and Turkish individuals with Broca’s aphasia indeed show deficiencies in the production and comprehension of verbs referring to the past, while verbs that refer to the future seem to be less difficult and verbs that refer to the present cause virtually no difficulties. Bastiaanse (2008) found that Dutch agrammatic speakers have more difficulty producing both finite and

(17)

non-17

finite verbs that refer to the past (both the third person singular past tense and participle) than their counterparts that refer to the present (third person singular present tense and infinitives). In 2011, Bastiaanse et al. conducted a study in which also for English and Turkish it was found that the production of verb forms referring to the past is impaired; more than referring to the present or future. However, for Chinese it was found that the production of all verb forms (past, present and future) was impaired. It was also found that in all three languages, comprehension of verbs that refer to the past was the most impaired. Their conclusion was that there is an underlying deficit in reference to the past through grammatical morphology in agrammatic aphasia. Yarbay Duman and Bastiaanse (2009) found similar data for agrammatic Turkish speakers: referring to the past through verbs seems to be the problem, and not tense. They present an epistemic modality model and assume that tense is a specific kind of epistemic modality marker that holds the distinction between [+/- past] and [+/- remoteness]. Verbs that refer to the past are remote ([+remote]) from the event, while present is not remote ([-remote]), and remoteness is not relevant in verbs that refer to the future, as the event has not happened yet; it does not exist (yet). Yarbay Duman and Bastiaanse propose that the nontemporal interpretation of epistemic modality is an issue when certainty of past (i.e. the certainty that an event has happened) is an issue. In other words: individuals with Broca’s aphasia have trouble with certainty that an event has happened, when there is no visual evidence available to support this certainty.

In the current study, this hypothesis will be investigated. Verbs that refer to the past inherently express epistemic modality (Yarbay Duman and Bastiaanse, 2009; Lyons, 1977): when one refers to the past through a verb form, one also expresses certainty that the event has happened. In these cases, there is often no visual evidence available; see the Introduction and section 2.1.4 for a discussion of visual evidence in past, present and future.

If individuals with Broca’s aphasia indeed have problems with the comprehension of the epistemic certainty value of verbs that refer to the past, they might also encounter difficulty with other words that express epistemic certainty when no visual evidence is available, like modal verbs.

Unfortunately, no extensive studies have yet been conducted on modal verbs or epistemic modality in aphasia. To our knowledge, there is only one small study by Yarbay Duman (in press) however, which tested 3 bilingual aphasia patients on “must” and “can”. The results showed that participants with aphasia had difficulty with epistemic inferencing. However, in child language acquisition research, modal verbs have been a more popular topic recently. In

(18)

18

the following section, several studies on the comprehension and production of modal verbs in typically developing children will be discussed.

2.2.1 Modal verbs in other populations

It has already been mentioned that there has been no extensive research yet on the production and comprehension of modal verbs in Broca’s aphasia. However, modal verbs have been a popular topic of research in typically developing (TD) children: among others Noveck (2001); Noveck, Ho and Sera (1996); Papafragou and Musolino (2003) and Ozturk and Papafragou (2015). Mostly, these studies were conducted to investigate the acquisition of scalar implicatures in TD children.

Noveck, Ho and Sera (1996) conducted their study to investigate the age at which English speaking TD children have an adultlike comprehension of the fact that different modal verbs form a scale that goes from strong to weak; thus, that have to expresses a stronger meaning than

might. They found that found that children of five years of age already have an understanding

of the logical meaning (see section 2.1.3) and use of the modal verbs have to and might: they systematically chose the stronger statement to be more certain than the weaker one, when put in contrast with each other. They also seem to understand the fact that a stronger statement does not necessarily means that the statement is more true, or more likely to be true. Seven-year olds already start to pick up on scalar implicatures; and nine-year-olds pick up on scalar implicatures as often as adults do.

This finding was supported by Noveck (2001), who conducted a study with five-year-olds, seven-year-olds and nine-year-olds who had English as native language. The aim of his study was also investigating the acquisition of scalar implicatures in children. This experiment included three boxes: two of them are open and can be looked inside, and one is covered. One of the open boxes contains a parrot and a bear, the other open box contains a parrot only. Participants were told that the covered box contains the same animals as one of the two open boxes. A puppet presents eight statements to the participants: (1) There has to be a parrot in

the box (true); (2) There does not have to be a parrot in the box (false); (3) There might be a parrot in the box (true); (4) There cannot be a parrot in the box (false); (5) There has to be a bear in the box (false); (6) There might be a bear in the box (true); (7) There does not have to be a bear in the box (true); (8) There cannot be a bear in the box (false). The participants had

(19)

five-year-19

olds and seven-year-olds accept might when must is appropriate (statement 3), while nine-year-olds fully reject this statement: showing a pragmatic interpretation.

Papafragou and Musolino (2003) did a study to see whether all weaker modal terms (like

some and the modal verb can) are treated the same by children. They tested 30 Greek children

with a mean age of 5;3 years old. They presented them with images which fitted a stronger statement (like all horses jumped over the fence), but described them with a weaker statement (like some horses jumped over the fence). They found that while almost all adults rejected these weaker statements, the children mostly accepted them. This supports the idea that five year olds are capable of understanding the logics of modal verbs, but scalar implicatures are not acquired yet.

Ozturk and Papafragou (2015) conducted three experiments to investigate the semantic understanding of modal verbs may and have to in English TD children of 4-5 years old, as well as the comprehension of scalar implicatures which arise when uttering a weaker modal when a stronger one is more appropriate. One of their experiments involved simple stories that provided cues about the location of the animal. There were two conditions: the statement condition, in which participants were presented modal statements about the location of the animal; and the question condition, in which participants were presented questions about the same topic. This division was made to rule out the possibility that young children would be more likely to agree with statements than with questions. Also, the participants were divided into two groups, of which one group would only hear items with may, and the other group would hear items with have to. The goal of this was not to give them a modal alternative, and that way encouraging the semantic interpretation. This approach appeared successful.

The participants were told they would play a game: there was an animal that was going to hide in a box. It could only hide in a box, and nowhere else. There were four story types in this study: one with one open box, in which the participants were to infer that the animal would hide in that box. The second story type included two boxes, and after the animal had hid itself, one of the boxes would open and appear empty; the test items then would target the open box; participants would have to infer that the animal has hidden in the other box. The third story type included two boxes as well, but both would stay closed after the animal had hidden. The test item targeted either one of the two closed boxes, which would both be a possible, but not necessary, hiding place for the animal. The fourth story type was identical to the second one, but the test item would target the closed box.

Again, it was found that children with a mean age of 5;1 years old understand the basics of epistemic modal semantics. They were the most accurate in the first story type, with one closed

(20)

20

box (96% accurate); in story types 2 and 4, with one open and one closed box, they were respectively 80% and 81% accurate; and in story type 3, in which the two boxes stayed closed, they were only 57% accurate. It appeared difficult for children to understand that it was only

possible for the animal to be in the targeted box, and not necessary. The children responded at

chance level in both the statement as the question group of the necessity condition and the question group of the possibility condition; they performed however above chance level (86%) in the statement group of the possibility condition. Children were generally more accurate in the statement groups, which suggests that the knowledge of the semantics of modal verbs has not yet completely developed at this age.

There have been some studies on modality in children with SLI (Specific Language Impairment) as well, but not on epistemic modality. Studies by Leonard et al. (2003) and Leonard, Deevy, Wong, Stokes and Fletcher (2007) only mention ability and permission (deontic and dynamic modality) in their study; and also do they only use the modal verb can. They showed that children with SLI seem to understand the notion of different kinds of modality, but have difficulty producing the correct forms.

The current hypothesis and study is therefore also relevant for children with SLI. Various studies have shown that children with SLI in many languages have difficulty with the production of past tense verbs (Oetting and Horohov, 1997; Van der Lely and Ullmann, 2001; Marshall and van der Lely, 2007; amongst others). Yarbay Duman and Topbas (in press) also have shown that children with SLI have difficulty with comprehending tense and aspect morphology; they suggest that non-temporal, epistemic functions of verb morphology may be relevant in the comprehension of tense and aspect morphology.

2.3 Summary and conclusion

In this chapter, both epistemic modality (2.1) and Broca’s aphasia (2.2) have been discussed. In the section on epistemic modality, the concept of modality has been explained, and the three types of modality (epistemic, deontic, and dynamic) as discussed by Nuyts (2001) have been introduced, to clarify the type of modality that is the topic of the current study: epistemic modality, which is the evaluation of the likeliness of a state of affairs to happen (or have happened) based on “what is known”. Then, different ways of expressing epistemic modality have been discussed. Modal auxiliaries are central in this study. In English, frequently used modal auxiliaries for expressing epistemic modality are might, can, must and have to. In Dutch,

(21)

21

the language that will be used in the test, frequently used modal verbs are kunnen and moeten: these will also be used in the test.

“What is known” determines whether someone expresses epistemic possibility (can or

kunnen) or epistemic certainty (must or moeten). What is known can be determined by, for

example, visual evidence (section 2.1.4): evidence that a speaker can see. There are two types of visual evidence: direct and indirect. Direct visual evidence is directly seeing an event happening (e.g. rain coming from the sky), while indirect visual evidence can cause a speaker to infer that the event has happened (e.g. wet streets that cause a speaker to infer that it has been raining). Available visual evidence, either direct or indirect, generally allows a speaker to express epistemic certainty.

In section 2.1.4.1, evidentiality is discussed, as different kinds of evidence are also very relevant in this concept: evidentiality is the expression of the source of information for a proposition. It has been mentioned in this chapter to make clear that, however they have similarities, evidentiality and epistemic modality are two different concepts.

When epistemic possibility is expressed while visual evidence is available, a scalar implicature (section 2.1.3) can occur. According to semantic theory, some means some and

possibly all, so a weak epistemic expression (like also can or kunnen) is acceptable in a situation

where epistemic evidence would allow for a stronger epistemic expression like must or moeten. However, to many people it will not appear appropriate: because when in a situation epistemic certainty is appropriate, why state a less certain expression? Many of the studies on the acquisition of modality and modal verbs in TD children are on the acquisition of scalar implicatures. A few of them are discussed in section 2.2.1.

Epistemic modality has not earlier been tested in individuals with Broca’s aphasia. Why it is relevant to do this has been discussed in section 2.2. Individuals with Broca’s aphasia often show difficulty in the production and comprehension of verbs that refer to the past (Bastiaanse, 2008; Yarbay Duman and Bastiaanse, 2009; Bastiaanse et al., 2011), and according to Yarbay Duman et al. (2011), verbs that refer to the past inherently express epistemic certainty. Other words that express epistemic certainty are discussed in section 2.1.1 and 2.1.2: must or moeten. It has been hypothesized that this expression of epistemic certainty causes problems in the comprehension of individuals with Broca’s aphasia, when no visual evidence is available. This hypothesis will be tested in the current thesis.

(22)

22

3 Hypothesis and predictions

In the current study, the idea by Yarbay Duman et al. (2011) was adopted: verbs that refer to the past, express epistemic certainty, as one is certain of the fact that the event that is being referred to has actually happened. Several studies showed that individuals with Broca’s aphasia often show problems in the production of verbs that refer to the past (Bastiaanse, 2008; Yarbay Duman and Bastiaanse, 2009; Bastiaanse, 2011; Yarbay Duman et al., 2011, among others). Yarbay Duman and Bastiaanse (2009) have hypothesized that individuals with Broca’s aphasia might have problems interpreting epistemic certainty, particularly when no visual evidence is available. This is the hypothesis that will be tested in the current study.

Fewer problems are found in the comprehension and production of verbs that refer to the future, where no epistemic certainty is possible, but visual evidence is often available. Even fewer problems are found with verbs that refer to the present: in this case, direct visual evidence is available, and epistemic certainty is therefore not relevant. This was explained in the Introduction and in section 2.1.4.

Consequently, the hypothesis for the current study is as follows:

“Dutch Individuals with Broca’s aphasia will have trouble comprehending the epistemic modal verb moeten, which expresses epistemic certainty; in particular when there is no visual evidence available to support this epistemic certainty.”

In the hypothesis, the modal verb moeten is explicitly mentioned as such; however, individuals with Broca’s aphasia are also expected to have cognitive problems with the notion of epistemic certainty. The problem is hypothesized to be a cognitive one rather than purely linguistic, as was found by Yarbay Duman (in press).

This hypothesis brings us to the following, more specific, predictions for the current test: 1. Individuals with Broca’s aphasia will have difficulty with the epistemic certainty modal

moeten when visual evidence is not available;

2. individuals with Broca’s aphasia will have no apparent difficulty with the epistemic certainty modal moeten when visual evidence is available;

3. individuals with Broca’s aphasia will have apparent difficulty with the epistemic possibility modal kunnen when visual evidence is not available;

(23)

23

4. individuals with Broca’s aphasia will have no apparent difficulty with the epistemic possibility modal kunnen when visual evidence is available;

5. individuals with Broca’s aphasia will have difficulty interpreting levels of epistemic certainty in picture stimuli – that is, when no epistemic modals are used.

Altogether this means that difficulty with epistemic modals is only to be expected when no visual evidence is available; and significantly more so when the modal auxiliary moeten is involved. See the table below for an overview.

Moeten Kunnen

Visual evidence available + + Visual evidence unavailable - +

Table 3: predictions for the performance of individuals with Broca’s aphasia in different situations; + means no problems expected, - means problems expected.

Also, individuals with Broca’s aphasia are expected to show difficulties with the comprehension of the notion of epistemic modality itself. Therefore, in situations where epistemic modality (i.e. visual evidence which can lead to varying levels of epistemic certainty) plays a role but no modal verbs are being uttered, difficulties with interpreting epistemic modality, or certainty, are expected to be found as well. If this is the case, and problems with interpreting the modal auxiliaries are found, this would support the findings by Yarbay Duman (in press) that the problems of individuals with Broca’s aphasia are indeed cognitive.

(24)

24

4 Methods

4.1 Participants

In total, twenty-six Dutch-speaking individuals participated in this study. Ten of them were individuals with Broca’s aphasia; seven of them were male and three were female. The other sixteen participants formed the control group of healthy adults; nine of them were female and six were male. The mean age of aphasia group was 56.1 years, and the mean age of the control group was 55.9 years; the groups were matched on age (t(df)=.042 p=.967; no significant difference). In the next section, the characteristics of the aphasia group will be discussed. After that, the control group of healthy adults and their characteristics will be discussed.

4.1.1 Aphasia group

The participants in the current study were approached through aphasia centers in Amsterdam, Rotterdam and surrounding areas. In these aphasia centers, patients learn to live with aphasia, and how to actively participate in daily life under professional guidance. All these individuals are patients with Broca’s aphasia, who are native speakers of Dutch and minimally 3 months post-onset.

All participants suffered at least one cerebrovascular accident (CVA) to the left hemisphere: one participant suffered more than one CVA (A10). Their ages varied from 36 to 76 years old (M=56.1; SD=12.8). Their post-onset time varied from 6 months to 174 months (M=58.3; SD=51.7). Participants had no articulation problems that might influence the test results, and they all had normal or corrected to normal vision and hearing. See table 4 for an overview of the characteristics of each individual with Broca’s aphasia.

The participants were selected by the speech therapists and clinical linguists of the aphasia centers, who knew the participants and their problems well. All participants were selected following the description of Bastiaanse and Thompson (2003) of Broca’s aphasia and had non-fluent and/or telegraphic output and word finding problems, but relatively spared comprehension. There were scores from the Dutch version of the Aachen Aphasia Test (AAT) available from most participants, to support the speech therapist’s selection to participate in this study; i.e. that they are an individual with Broca’s aphasia (see appendix A for the scores of the participants on the AAT that were available, and the next section for more information about the AAT).

(25)

25

Gender Age Hand. Education Profession Monolingual MPO AAT Etiology

A1 F 42 R Academic Program manager Yes 87 No LCVA

A2 M 44 R Vocational None Yes 13 Yes* LCVA

A3 M 61 R Technical Technical manager Yes 48 Yes LCVA

A4 F 76 R Vocational Shop assistant Yes 6 Yes* LCVA

A5 M 70 R Vocational Public relations Yes 41 Yes* LCVA

A6 F 36 R Vocational Beautician Yes 12 Yes LCVA

A7 M 55 R Vocational Entrepreneur Yes 60 Yes LCVA

A8 M 54 R Vocational Actor/director Yes 22 Yes* LCVA

A9 M 66 R Vocational Banker Yes 84 Yes* LCVA

A10 M 57 R Vocational Superintendant Yes 174 No LCVA

Table 4: characteristics of Aphasia Participant (A) 1 to 10, the aphasia group; F: Female; M: Male; R: Right; MPO: Months Post-Onset; AAT: data available? Yes*=partly; LCVA: Left CVA.

4.1.1.1 Aachen Aphasia Test (AAT)

AAT scores are available for most participants in the aphasia group. The Aachen Aphasia Test (AAT; Huber et al., 1983) is an originally German test for assessing the severity and type of aphasia. It has been translated to Dutch by Graetz, De Bleser and Willmes (1992).

The AAT consists of six subtests: spontaneous speech, a token test, repeating, writing, confrontation naming, and comprehension. See table 5 for an overview of the subtests and the scoring.

Name of subtest Scoring: Parts: Spontaneous

Speech

0 - 30 (1) Communicative behavior; (2) Articulation and prosody; (3) automatic speech; (4) semantic structure; (5) phonemic structure; (6) syntactic structure.

All subtests could be scored 0 (heavily deficient) to 5 (not deficient)

Token Test 0 - 50 50 items; all subtests could be scored 0 (incorrect) or 1 (correct)

Repeating 0 - 150 (1) sound repetition; (2) monosyllables; (3) loanwords; (4) compound words; (5) sentences

Writing 0 - 90 (1) Reading out loud; (2) showing the letters of a dictated word off a page; (3) writing down a dictated word

(26)

26 Confrontation

naming

0 - 150 (1) naming simple nouns; (2) naming colors; (3) naming compound nouns; (4) describe situations and actions through a sentence

Comprehension 0 - 150 (1) auditory word comprehension; (2) auditory sentence comprehension; (3) word comprehension of written text; (4) sentence comprehension of written text.

Table 5: subtests and possible scoring of the Aachen Aphasia Test.

4.1.2 Control group

The control group consisted of 16 native speakers of Dutch. None of the adults had a history of speech and/or language disorders. These adults were approached online and via personal connections. The control group participants varied from 45 to 66 years of age (M=55.9; SD=7.1). See table 6 for the characteristics of the individuals in the control group.

Table 6: characteristics of Control (C) 1 to 16, the control group; Hand: handedness; F: Female; M: Male; R: Right. In the case of no current profession, previous professions are not known by the experimenter.

Gender Age Hand Education Profession Monolingual

C1 F 46 R Academic None Yes

C2 F 45 R Academic Change manager Yes

C3 F 50 R Academic Editor Yes

C4 F 49 R Vocational Tour guide Yes

C5 F 54 R Vocational Instructor Yes

C6 M 49 R Vocational Accountmanager Yes

C7 F 51 R Vocational None Yes

C8 M 65 R Vocational None Yes

C9 F 66 R Vocational Nutritionist Yes

C10 M 60 R Vocational None Yes

C11 M 52 R Academic Advisor Yes

C12 M 63 R Vocational None Yes

C13 F 61 R Vocational Hotel director Yes C14 M 60 R Vocational Entrepreneur Yes

C15 F 61 R Vocational Advisor Yes

(27)

27 4.2 Experimental design

The current test exists of two tasks: (1) a sentence judgment task in which participants are asked to respond to statements about the location of an animal; and (2) an epistemic inferencing task, called the pointing task, in which they are asked to answer questions about their level of certainty about the location of an animal in different situations without linguistic cues. These questions are answered by pointing to the screen, hence the name pointing task. The first task is an adaption of the test by Ozturk and Papafragou (2015), which has been discussed in section 2.2.1, but question task was added to answer to the fifth prediction mentioned in the previous section: to make sure individuals with Broca’s aphasia are able to differentiate between different levels of epistemic certainty when no words are being used to express this. The same test was administered to all participants.

The participants were presented with short stories on a laptop screen, which provided cues about the location of an animal. The participants were told that animals would appear on the stage on screen, and that they would hide in a box.

The sentence comprehension task consisted of two conditions: the kunnen condition (epistemic possibility) and the moeten condition (epistemic certainty). Each item for the sentence comprehension task includes at least three slides. During the last slide, a sentence is presented to the participant. The sentences are a statement that contain a modal verb (kunnen or moeten, depending on the condition) and a proposition about the location of an animal, which was always in a box (e.g. de kat moet in de bruine doos zitten; ‘the cat must be in the brown box’). Participants were asked to respond to these statements with yes or no. See figure 1 for an example of slides that are being shown in the sentence judgment task. This task also included four different story types, which included different kinds of evidence. They will be explained more extensively below.

(28)

28

The aim of this first task is to test the comprehension of Dutch modal verbs kunnen and

moeten of the aphasia group and control group: relevant to the first four predictions that were

mentioned in the previous section. There were two conditions with statements: the kunnen condition and the moeten condition. Participants were asked to respond with either yes or no in these conditions, which included modal verbs.

The second task is aimed at testing the abilities of epistemic inferencing when no words that express epistemic certainty are involved: relevant to the fifth prediction mentioned in the previous section. In this task, called the pointing task, three images were shown to the participant on one slide (see figure 2). In this task, the participants were asked to point to the picture out of the three presented on the screen, on which they were most certain that the animal was in the box that was closed in all three figures; i.e. the brown box in figure 2.

Figure 2: example of visual stimuli in the pointing task

In total, this test included three components: (1) kunnen, (2) moeten (both part of the sentence comprehension task), and (3) the pointing task. There were 48 stimuli per participant: 2 conditions (kunnen and moeten) in the sentence judgment task with each 16 stimuli; the pointing task with 4 stimuli; and also 12 filler items. See table 7 below for an overview of the conditions and their story types; see also table 9 at the end of this section for a complete overview of the conditions, their story types, examples, targets, and predictions.

(29)

29

Task Sentence comprehension task Pointing task

Condition Moeten Kunnen

Story type 1. One closed box 1. One closed box (none)

2. One closed, one open; statement about open box

2. One closed, one open; statement about open box 3. One closed, one open;

statement about closed box

3. One closed, one open; statement about closed box

4. Two closed boxes 4. Two closed boxes

Table 7: An overview of the conditions and their story types.

The kunnen and moeten condition in the sentence comprehension task each had four story types ((1) one closed box; (2) one closed, one open; statement about open box; (3) one closed,

one open; statement about closed box; and (4) two closed boxes); these story types were the

same as the four story types in Ozturk and Papafragou (2015). See table 5 for an overview of the conditions and their story types. All four story types of the kunnen and moeten conditions were repeated four times with different animals and boxes. The pointing task had no different story types: this condition was also repeated four times with different boxes and animals. Furthermore, there were twelve filler items in total. See table 8 for an overview of the number of items for each condition. Before starting the test, four test items were shown to support the instructions.

Task/cond Var.

Moeten Kunnen Pointing Filler

1 4 items 4 items 4 items 12

2 4 items 4 items

3 4 items 4 items

4 4 items 4 items

Total 16 16 4 12 48

Table 8: Overview of the amount of items for each condition.

Four combinations of an animal with boxes were used in the test: each repetition of a test item, another combination was shown. The combinations were the following: a monkey with a

(30)

30

grey and a pink box; a rabbit with a brown and a purple box; a lion with a green and a white box; and a dog with a red and a black box. See figure 1 for all the animals and boxes that were used. Another combination (of a cat with a purple and brown box) was used in the practice items; this combination will also be used in the examples below in section 5.2.1.

The background story in the test was the same as in Ozturk and Papafragou (2015): an animal will be appear on stage, and it is going to hide itself. t will always hide in a box (so not behind the curtains or somewhere off stage).

In the sentence comprehension task, a stage was presented with an animal on it, as well as one or two boxes. The next slide shows closed curtains. When the curtains open again in the slideshow, the animal has gone (it is hidden), and a statement (which included kunnen in the kunnen condition, and moeten in the moeten condition) was read out loud by a native speaker of Dutch. The participants were asked to respond with either yes or no. In the next section, the conditions will be explained further and examples will be presented.

In the pointing task, one slide with three images was shown: one of the images presented one closed box; another presented an open and a closed box; and the last one two closed boxes (see figure 8 in the next section for an example) for examples of these on a stage. The experimenter asked in which image the participants were most certain about the location of the animal; the participants were asked to respond by pointing to the image on the screen.

The same materials were used for both groups. In the following section, all details of the sentence comprehension task and its two conditions and four story types will be discussed; subsequently, the details of the pointing task will be explained.

(31)

31

Figure 3: the animal and box combinations used in the test items.

4.2.1 Conditions, story types and filler items

It has been mentioned that both the kunnen and moeten conditions each have four different story types (see table 7 for an overview). These four story types are different in the amount of (open and/or closed) boxes shown on stage and which box the statement is targeted on.

The four story types are identical in both the kunnen and moeten condition, and only differ in the use of the modal auxiliary in each condition. The examples given here include a cat with a purple and a brown box, which was the animal-box-combination used in the practice items. The test items mentioned here as examples also include this cat and purple and brown box; in the test, the test items obviously are in agreement with the animal-and-boxes-combination displayed at that moment.

The first story type is the one closed box story type. In figure 4, the sequence of slides in this story type can be found.

Dutch individuals with Broca’s aphasia and their comprehension of epistemic (un)certainty with or without visual evidence