Imageability vs. Concreteness: the same or different?
Student: Natalie Koetsoeba
S3707520
University of Groningen
MA Neurolinguistics
Thesis supervisor: Dörte de Kok
Date: 01-07-2020
Index
1. Introduction ... 3
1.1 Language processing in healthy individuals at the level of words ... 3
1.2 Psycholinguistic variables ... 4
1.3 Word retrieval in people with aphasia ... 6
1.4 Imageability and Concreteness ... 7
1.5 The current study ... 8
2. Method ... 9
2.1 New imageability ratings ... 9
2.1.1 Participants ... 9
2.1.2 Materials and procedure ... 10
2.1.3 Analyses ... 10
2.2 Naming task ... 10
2.2.1 Participants ... 10
2.2.2 Materials and procedure ... 11
2.2.3 Analyses ... 11
3. Results ... 12
3.1 New imageability ratings ... 12
3.1.1 Descriptive statistics questionnaire ... 12
3.1.2 Correlation analysis ... 12
3.2 Naming task ... 13
3.2.1 Descriptive statistics naming task ... 13
3.2.2 Regression analysis ... 14
4. Discussion ... 15
4.1 Coherence variables ... 16
4.2 New imageability vs. old imageability (Van Loon-Vervoorn, 1985) ... 17
4.3 Imageability vs. Concreteness ... 17
4.4 Limitations and future research ... 18
5. Conclusion ... 19
References ... 20
Appendices ... 25
Appendix A – Newly acquired imageability ratings ... 26
Appendix B – Characteristics participant group naming task ... 28
Appendix C – Naming task ... 29
1. Introduction
Aphasia is “an acquired language disorder, caused by focal brain injury that occurs after the language is acquired” (Bastiaanse, 2011, p. 12). Possible causes of aphasia can be a stroke, trauma, tumor or an inflammation, where stroke is the most common cause of aphasia; approximately 30% of the people who experience a stroke acquire aphasia (Engelter et al., 2006; Laska, Hellblom, Murray, Kahan, & Von Arbin, 2001). In the Netherlands the prevalence of people with aphasia is estimated at approximately 30.000, with around 10.000 new diagnoses each year (Bastiaanse, 2011). People with aphasia can experience various difficulties of varying severity and therefore form a heterogenous group (i.e. no person with aphasia is the same). They can experience difficulties in various language modalities, such as language production and comprehension, reading and writing. These difficulties can be relatively mild, for example word-finding difficulties, or global, where all modalities are severely impaired and a person is hardly able to speak (Code & Herrmann, 2003). In the current study the focus will be specifically on difficulties at the level of words.
1.1 Language processing in healthy individuals at the level of words
In 1988, Ellis and Young proposed an extensive model for language processing at the level of words (see figure 1). With the help of this model, disorders in specific components of language processing can be localized and identified. A disorder in a specific component can cause, for example, word retrieval deficits or difficulties with finding the correct phonemes of a word, which can restrict a person with aphasia in his language and communication. The Psycholinguistic Assessment of Language Processing in Aphasia (PALPA; Kay, Lesser, & Coltheart, 1992, Dutch version: Bastiaanse, Bosje, & Visch-Brink, 1995) is a comprehensive test battery for underlying disorders in people with aphasia, consisting of 52 tasks in which each test investigates the functioning of a certain component or route in the model of Ellis and Young (1988).
Figure 1. Model for language processing at the level of words (Brunsdon, Coltheart, & Nickels, 2005; Ellis & Young, 1988).
Auditory recognition
The Auditory Analysis System distinguishes, recognizes and orders the individual phonemes in a speech wave. Subsequently the phonological word form is activated in the Phonological Input Lexicon, provided that it is a familiar word. Words that are phonologically related (e.g. ‘cat’ and ‘rat’) are stored close together and are also coactivated. At this stage no meaning is assigned to the set of phonemes heard. The activation of the meaning of the heard word takes place in the Semantic System. In this
system, lemmas are stored and ordered by meaning. For example, ‘cat’, ‘dog’ and ‘horse’ are stored close together. In the Semantic System there is also coactivation. In a healthy language system, the activation for the heard word is the highest and the coactivated words will extinguish (Bastiaanse, 2011; Ellis & Young, 2014).
Oral language production
The production of a spoken word starts with the activation of a certain concept in the Semantic System. Hence a concept can consist of many semantic features, there are several words that belong to one concept (Boyle, 2010). For example, the concept ‘a tree in autumn colors’ needs only one thought but more than one word to convey the idea (Links, Feiken, & Bastiaanse, 1996). The associated word form is subsequently activated in the Phonological Output Lexicon. In the Phonological Output Lexicon all spoken word forms are stored. Similar to the Phonological Input Lexicon, words that are phonologically related are stored closely together and are also coactivated. In a healthy language system, the coactivated words will extinguish and the correct word will win. The final step is assigning the correct phonemes in the correct order to the word form activated in the Phonological Output Lexicon. This process takes place in the Phonological Output Buffer (Bastiaanse, 2011; Feiken, Hüttmann, & Links, 2020).
Visual recognition
The process of recognizing written words is comparable to that of auditory words. The Visual Analysis System distinguishes, recognizes and orders the individual graphemes and after this the written word form is activated in the Orthographic Input Lexicon. Only familiar words are stored here, the same as for the Phonological Input Lexicon. Thereafter in the Semantic System a certain concept is retrieved and the meaning of the word is assigned to the written word recognized in the Orthographic Input Lexicon (Bastiaanse, 2011; Ellis & Young, 2014).
Written language production
The production of a written word proceeds in a similar way as the production of a spoken word. It starts with the activation of a certain concept in the Semantic System. Subsequently the corresponding word form is activated in the Graphemic Output Lexicon. Thereafter the correct graphemes are assigned in the correct order to the word form, which takes place in the Graphemic Output Buffer (Bastiaanse, 2011).
Sublexical processing
As shown in the model, there are also sublexical routes that go beyond the specific lexicons. Using these routes, it is possible to repeat, read out loud or write for example non words. Non words are defined as sound and letter sequences that do not have a lexical representation (Links et al., 1996).
1.2 Psycholinguistic variables
Psycholinguistic variables describe the properties of a word, such as how often a word occurs or at what age a word is acquired. These variables are used to learn more about language processing in people with language disorders, such as aphasia, but can also be measured in healthy individuals. These variables are indicative of impairments in a specific component, looking at the different components in the model of Ellis and Young (1988), see figure 2. The processing of words is dependent on these variables, which in people with aphasia is particular visible in the performance on language tasks (Bastiaanse, 2011) and in healthy individuals in the reaction times on language tasks (Gernsbacher, 1984; Paivio, 1991). Examples of psycholinguistic variables are Frequency, Age of Acquisition, Length, Imageability and Concreteness.
Frequency indicates how often a certain word occurs, which can be determined by counting the occurrences of a word in a finite specimen of text that is thought to be representative of language as
a whole (Keuleers, Brysbaert, & New, 2010). Age of Acquisition means the age at which a word is acquired. Age of Acquisition ratings are collected by asking participants the age at which they thought they had learned a specific word. More specifically, the age at which they would have understood the word when it was used by someone else, even if they did not use the word themselves (Bysbaert, Stevens, De Deyne, Voorspoels, & Storms, 2014). Frequency and Age of Acquisition both give information about the Output Lexicon, but they still complement each other (Rofes, de Aguiar, & Miceli, 2015). Frequency is the most influential variable for lexical decision tasks and explains easily 30% of the variance in lexical decision times, when obtained from an adequate corpus (Balota et al., 2007; Brysbaert & New, 2009; Ferrand et al., 2011, 2010; Keuleers, Diependaele, & Brysbaert, 2010; Keuleers, Lacey, Rastle, & Brysbaert, 2012). However, Frequency does not fully correspond to the cumulative frequency with which participants have been exposed to words (Bonin, Barry, Méot, & Chalard, 2004; Pérez, 2007; Zevin & Seidenberg, 2002). Kuperman, Stadthagen-Gonzalez and Brysbaert (2012) found that Age of Acquisition explains 2-10% of the variance in lexical decision times above the effects of Frequency. Besides, Age of Acquisition gives information about the order in which words are learned, which influences the speed with which their representations can be activated, independently of how frequent a word is (Kuperman et al., 2012). Although we assume in the current study that Age of Acquisition gives information about the Output Lexicon, there are considerable findings that Age of Acquisition also gives information about the Semantic System. For example, Van Loon-Vervoorn (1989) used a discrete word-associate generation task to tap into the Semantic System, where participants were asked to name the first word that came to their mind when seeing a stimulus word. They found a reliable effect for Age of Acquisition and no effect for Frequency (Brysbaert, Van Wijnendaele, & De Deyne, 2000). On the other hand, Morrison, Ellis and Quinlan (1992) failed to find an effect for Age of Acquisition in a semantic task where participants had to classify pictures of objects as naturally-occurring or man-made.
Length gives information about the number of phonemes of a word and gives information about the buffer (Rofes et al., 2015). Imageability gives an indication of how well a word gives rise to a mental image (Carroll & White, 1973). Finally, Concreteness gives an indication of the degree to which a concept denoted by a word refers to a perceptible entity (Bysbaert et al., 2014). Imageability and Concreteness both give information about the semantics (Rofes et al., 2015). The current study will focus specifically on the psycholinguistic variables Imageability and Concreteness, since these two are often used interchangeably in the literature (Reilly & Kean, 2007; Tyler & Moss, 1997; Tyler, Moss, Galpin, & Voice, 2002).
1.3 Word retrieval in people with aphasia
One of the most common impairments that people with aphasia experience are word retrieval deficits (Howard & Gatehouse, 2006; Nickels, 1997). These deficits are present in a picture naming task, but also in spontaneous speech. In spontaneous speech, these deficits result in empty speech, where a person with aphasia speaks fluently but explicit information, like nouns, verbs and adjectives, is missing. Also, hesitations, hitches and paraphasias can occur (Bastiaanse, 2011). Looking at the model of Ellis and Young (1988), these deficits can be caused by an impairment in the Semantic System (Hillis, Rapp, Romani, & Caramazza, 1990), an impairment in accessing the Phonological Output Lexicon from the Semantic System or in the Phonological Output Lexicon itself (Kay & Ellis, 1987). There can also occur retrieval problems in the written modality caused by an impairment in the Orthographic Output Lexicon (Rosati & De Bastiani, 1979).
As mentioned, one cause of word retrieval deficits can be caused by an impairment in the Semantic System. This is also known as a central semantic deficit. When there is a central semantic deficit, a person with aphasia is impaired in both comprehension and production, and across all four modalities (Bastiaanse, 1991, 2011; Butterworth, Howard, & Mcloughlin, 1984). People with a semantic deficit will perform poor on comprehension tasks, will make semantic errors in production, and are poor in detecting those errors. These semantic errors do occur in naming tasks as well in tasks which do not require spoken responses (Ellis & Young, 2014). They show often category-specific disorders, where a person is able to name objects in some semantic categories but not others (e.g. Hart, Berndt, & Caramazza, 1985; Warrington & Shallice, 1984). Also, an imageability effect can occur, which means words with a low imageability are more difficult to retrieve than words with a high imageability (Bastiaanse, 2011).
The cause of the word retrieval deficits can also be a deficit in the access to the Output Lexicon from the Semantic System. When the deficit is in the access between these two modules, the words meanings are available in detail but not the word itself. Repetition of existing words is intact and there is no consistent error pattern (Kay & Ellis, 1987). Other characteristics of an impairment in the access to the Output Lexicon from the Semantic System are word retrieval deficits, semantic paraphasia's (Hashimoto, Widman, Kiran, & Richards, 2013) and an imageability effect (Bastiaanse, 2011; Howard & Gatehouse, 2006). Furthermore, high frequency words are often easier to retrieve than low frequency words. This is called a frequency effect (Bastiaanse, 2011).
Finally, word retrieval deficits can also be caused by an impairment in the Lexicon itself. When there is an impairment in the Phonological Output Lexicon or in the Orthographic Output Lexicon, a person with aphasia is not able to activate the correct wordform (Biran & Friedmann, 2005). Sometimes a person does have partial access to a wordform, which results in phonological approximations, phonological paraphasias or neologisms (Dorze & Nespoulous, 1989; Ellis & Young, 2014; Kay & Ellis, 1987). Also, frequency effects occur in this type of impairment (Bastiaanse, 2011). Specific for an impairment in the Orthographic Output Lexicon are difficulties with the spelling of irregularly spelled words (Bastiaanse, 1991, 2011). However, in clinical practice many people with aphasia who experience word retrieval deficits show a combination of impairments in the Semantic System and the Lexicon, because their brain injury affects both levels (Ellis & Young, 2014; Howard & Gatehouse, 2006).
Various studies have studied the influence of the psycholinguistic variables on word retrieval in people with aphasia, but also in healthy individuals. Nickels and Howard (1995) investigated which psycholinguistic variables could affect the naming performance with the help of an oral naming task, in thirteen people with aphasia. They found that the naming performance of many participants was significantly affected by Age of Acquisition. Imageability, Concreteness and Length were also predictive for the naming performance of some participants. For Frequency they found a small effect. However,
the participants showed different patterns on the various psycholinguistic variables and there was a wide variation between them. Bastiaanse, Wieling and Wolthuis (2016) studied the same in 54 people with aphasia. They found that Imageability, Frequency and Age of Acquisition play a significant role in the retrieval of nouns. For healthy individuals, Morrison and Ellis (1995) found significant effects of Age of Acquisition on immediate word-naming speed in a word-naming task, where the participants had to read aloud words presented to them on a computer screen. Spieler and Balota (2000) conducted a similar task in healthy individuals and found significant effects for Frequency and Length. Since the current study will focus specifically on Imageability and Concreteness, a more detailed explanation of these two variables follows below.
1.4 Imageability and Concreteness
As described earlier, Imageability (also named Imagery) is the psycholinguistic variable that gives an indication of how well a word gives rise to a mental image (Rofes et al., 2018; Van Loon-Vervoorn, 1985). For example, the word ‘house’ has a high imageability and the word ‘fact’ has a low imageability. It has been found that high imageability words are processed faster and more accurately than words with a low imageability. This is called the imageability effect (Bastiaanse, 2011; Howard & Gatehouse, 2006; Rofes et al., 2018). Strain et al. (1995) found this imageability effect in healthy individuals performing a word naming task. Alario et al. (2004) found an imageability effect in healthy individuals performing a picture naming task, which was particularly notable given the restricted range of imageability values in their study. The words they used in their study were all very high imageable. Nickels and Howard (1995) found this imageability effect also in people with aphasia. They found that some people with aphasia were better at naming high imageable pictures than low imageable pictures, even after controlling for the effects of other variables. Howard & Gatehouse (2006) found the same result in two people with aphasia performing a picture naming task. The ratings for imageability values are typically collected through questionnaires, where one has to rate how imageable certain words are. The instruction is usually as follows. Participants are asked to rate a list of words as to the ease or difficulty with which the words arouse mental images. A word which arouses a mental image very quickly and easily should be given a high imageability rating and a word which arouses a mental image with difficulty or not at all should be given a low imageability rating (Paivio, Yuille, & Madigan, 1968). There are many databases with imageability ratings available in different languages such as English (Juhasz, Lai, & Woodcock, 2014), Portuguese (Soares, Costa, Machado, Comesaña, & Oliveira, 2017), Chinese (Yao, Wu, Zhang, & Wang, 2017), Polish (Imbir, 2016), Italian (Vergallito, Petilli, & Marelli, 2020), French (Chedid et al., 2019) and Dutch (Van Loon-Vervoorn, 1985). When looking at all these databases it is striking that most of these databases are very recent, except the database for the Dutch language. Compared to the other databases, the database of Van Loon-Vervoorn (1985) is a very old and rather small database which is not available online. This could mean the imageability ratings for the Dutch language are outdated and furthermore it is more time consuming when searching for the imageability rating for a specific word, since the database is not available online and the rating for a specific word has to be looked up manually.
Concreteness is the psycholinguistic variable that gives an indication of the degree to which a concept denoted by a word refers to a perceptible entity (Bysbaert et al., 2014). According to Paivio (2013) concrete words are easier to recall than abstract words. This is called the concreteness effect. Sandberg and Kiran (2014) found that both healthy individuals and persons with aphasia have faster reaction times and more accurate responses for concrete words compared to abstract words on a word judgment task and a synonym judgment task. In the word judgment task, the participants had to decide whether a word was abstract or concrete. In the synonym judgment task, the participants had to decide whether or not two abstract words and two concrete words were similar in meaning. Besides, they found an exaggerated concreteness effect in the persons with aphasia, assuming that the lesion may influence the neural correlates of abstract and concrete word processing. Roxbury, McMahon and Copland (2014) and Kounios and Holcomb (1994) both found a concreteness effect in healthy
individuals performing a lexical decision task, where participants had to decide whether a word exists or not. They both found that healthy individuals have significantly faster reaction times for concrete words compared to abstract words. People with aphasia also show a better performance for concrete words than abstract words on tasks for word repetition (Martin, Saffran, & Dell, 1996), word reading (Newton & Barry, 1997), word recognition (Crutch & Warrington, 2005) and reading comprehension (Barry & Gerhand, 2003). The ratings for concreteness values are typically collected in the same way as imageability values, namely through questionnaires. In these questionnaires a participant has to rate if a word is abstract or concrete. The instruction is usually as follows. Participants are asked to rate a list of words and to indicate per word how concrete or abstract the word is. A concrete word refers to something that exists in reality and it can be experienced with your senses or by taking an action. An abstract word refers to something that cannot be directly experienced (Bysbaert et al., 2014). There are several databases with concreteness ratings available in different languages such as English (Brysbaert, Warriner, & Kuperman, 2014), French (Bonin, Méot, & Bugaiska, 2018), Portuguese (Soares et al., 2017) and Chinese (Yao et al., 2017). In contrast to Imageability, there is a recent database available for Concreteness for the Dutch language (Bysbaert et al., 2014).
Imageability and Concreteness are two nominal variables that appear to be highly correlated with each other. Correlations found in the literature range from .78 to .85 (Friendly, Franklin, Hoffman, & Rubin, 1982; Gilhooly & Logie, 1980; Paivio, 2013; Paivio et al., 1968). Despite these two psycholinguistic variables being highly correlated, they are not the same. In general, it is the case that words with a low concreteness also have a low imageability (e.g., fact) and words with a high concreteness have a high imageability (e.g., apple). However, Paivio et al. (1968)also found words that are higher in Imageability than Concreteness (e.g., affection, blessing, ghost, delirium and hierarchy) and the other way around (e.g., antitoxin, encephalon, originator). In addition, the literature is divided on whether Imageability or Concreteness is a better predictor. Connell and Lynott (2012) preferred Concreteness over Imageability, since Imageability would stress the visual modality too much.Guasch, Ferré and Fraga (2016) found that Concreteness, and not Imageability, was significantly predictive for reaction times on a lexical decision task in healthy Spanish speakers. Furthermore, Concreteness and Imageability appeared to be both significant predictors of reaction times in a lexical decision task involving Chinese characters, but Concreteness accounted for 8% of the variances in the reaction times and Imageability only for 1% (Yao et al., 2017). On the other hand, Marcel & Patterson (1978) preferred Imageability over Concreteness, as they found that visual word recognition was affected by Imageability and not by Concreteness, for both healthy individuals and people with aphasia. Imageability would also be a significantly better predictor of memory scores, compared to Concreteness (Paivio, 1967). Dellantonio, Mulatti, Pastore and Job (2014) showed that Imageability and Concreteness do not measure the same and that Imageability is the more interesting variable. Imageability does not only reflect the ease with which words evoke external sensations, but also reflects the ease with which words evoke body-internal sensations. However, although the literature is very clear that those two variables are not the same, many researchers have used the two terms Imageability and Concreteness interchangeably (Dellantonio et al., 2014; Reilly & Kean, 2007; Tyler & Moss, 1997; Tyler et al., 2002).
1.5 The current study
As far as we know, there has been no study yet that investigated specifically the two psycholinguistic variables Imageability and Concreteness in word retrieval. However, often one of the two psycholinguistic variables is included in word retrieval studies without knowing which one is the better predictor for the outcomes on that task.
The current study is going to investigate the effect of Imageability and Concreteness in healthy individuals. A naming task will be administered and the predictive value of the imageability and concreteness ratings for the reaction times will be evaluated. This is important in order to clarify the role of these two variables in word retrieval and to underline that these variables cannot be used
interchangeably. Besides, this study can contribute to a better understanding of the psycholinguistic variables involved in naming and word retrieval in healthy individuals. This leads to the following research question and subquestions:
What is the influence of the psycholinguistic variables Imageability and Concreteness on word retrieval for healthy individuals?
- How strong is the correlation between Imageability and Concreteness?
- Are the imageability ratings of Van Loon-Vervoorn (1985) still valid or are they due for replacement?
- Does Concreteness or either of the two imageability ratings (Van Loon-Vervoorn (1985) vs. newly acquired ratings) have a higher predictive value for the reaction times in a word retrieval task than the other measures?
Based on the literature, the hypothesis is that there is a difference in the processing of the psycholinguistic variables Imageability and Concreteness for healthy individuals. There is expected that healthy individuals have faster reaction times for the words with a high imageability and for concrete words (Paivio, 1991; Sandberg & Kiran, 2014; Strain et al., 1995). Furthermore, a strong correlation between the variables Imageability and Concreteness is expected (Friendly et al., 1982; Gilhooly & Logie, 1980; Paivio, 2013; Paivio et al., 1968). Also, a strong correlation between the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings in the current study is expected, because the ratings are about the same psycholinguistic variable and collected in a similar way. Finally, it is expected that Concreteness can best explain the reaction times, since there is more evidence in the literature that Concreteness significantly predicts reaction times (Guasch et al., 2016; Yao et al., 2017).
2. Method
The main aim of the current study is to investigate the effect of Imageability and Concreteness in healthy individuals. To investigate this, a digital written naming task was administered to the participants, measuring the reaction times. Furthermore, new imageability ratings were acquired to see if the imageability ratings of Van Loon-Vervoorn (1985) are still valid. First, an elaboration of the method of the newly acquired imageability ratings follows and thereafter the elaboration of the naming task.
2.1 New imageability ratings
2.1.1 Participants
The participant group that participated on the questionnaire for the new imageability ratings consisted of 81 participants, fourteen men and 66 women. The gender of one participant was unknown. The participants were recruited through the Linguistics department of the University of Groningen, Social Media and the social environment of the researcher. All participants were informed about the aims of this study and have given their permission (informed consent) to this. The inclusion criteria were as follows:
Inclusion criteria:
- A minimum age of 18 years old; - Dutch as a native language.
The mean age of the participants was 27;9 years (range 18-65). Three participants were excluded, since their native language was German.
2.1.2 Materials and procedure
A questionnaire was drawn up on the website www.qualtrics.com. The questionnaire consisted of the fifty words who belonged to the descriptions of the naming task. The participants were asked to give an indication of how easy it would be to make a drawing of a certain word. It was mentioned that this was not about the drawing talent of the participant, but the extent that the word evokes a mental image. Also, an example was mentioned: ‘for example, the word "chair" is easier to draw than the word "thought”’. The participants could always choose from five categories: 1 = very easy, 2 = easy, 3 = average, 4 = difficult, and 5 = very difficult. There was also an option for the participants to indicate that they did not know the word. There has been chosen for a 5-point Likert scale, because this is less confusing for the participants and increases response rate and response quality (Babakus & Mangold, 1992; Devlin, Dong, & Brown, 2003; Hayes, 1992). To control for any effects as a consequence of the sequence of the words, the order of the words was randomized for all participants.
2.1.3 Analyses
The mean imageability ratings, the standard deviations and the number of times a word was unknown for the participants, were calculated for each word (see Appendix A). Thereafter, the imageability ratings of Van Loon-Vervoorn (1985), the concreteness ratings of Brysbaert, et al. (2014) and the newly acquired imageability ratings have been converted to z-scores with the help of SPSS (version 26). The reason for this is that all three values had different scales. Finally, a correlation analysis has been conducted in SPSS to determine the correlation between the concreteness ratings of Brysbaert et al. (2014), the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings. A Spearman’s rank correlation was chosen, because the three variables are on an ordinal level. Results were interpreted based on the coefficient Spearman Rho. A very weak correlation is between 0.00 – 0.30, a weak correlation is between 0.30 – 0.50, an average correlation is between 0.50 – 0.70, a strong correlation is between 0.70 – 0.90 and a very strong correlation is between 0.90 – 1.00. Significance was met at 𝝰-level 0.05.
2.2 Naming task
2.2.1 Participants
The participant group that participated on the naming task consisted of 51 healthy participants, thirteen men and 38 women. All the participants were recruited throughout the southern and northern parts of the Netherlands, through the social environment of the researcher. All participants were informed about the aims of this study and have given their permission (informed consent) to this. The inclusion criteria were as follows:
Inclusion criteria:
- A minimum age of 18 years old; - No history of neurological damage;
- No impairments that make it impossible to participate in the current study1.
The mean age of the participants was 30 years (range 19-65). 43 participants were right-handed and eight were left-handed. None of the participants had a history with neurological damage. The classification by highest level of education is based on the Standaard Onderwijsindeling 2016, Editie 2018/’19 (Centraal Bureau voor de Statistiek, 2019). Based on this classification there were eleven participants with a medium level of education and 32 with a high level of education. For a more detailed description of the characteristics of the participant group, see Appendix B.
1Note. Only native speakers of Dutch have been approached to participate in the current study. Most likely,
2.2.2 Materials and procedure
The naming task that has been used is based on a naming task that Voshart (2019) developed for her master’s thesis. See Appendix C for an overview of the items in the naming task. This naming task has been converted in a digital version with the tool PsychoPy (Peirce et al., 2019) and run on
www.pavlovia.org. The test consisted of fifty descriptions. The participants were asked to fill in the
word matching the description, using one word, using only nouns and do this as accurately and quickly as possible. There were two example descriptions to which the participants received feedback, to make sure that the participants understood the intention. The examples were two items with a high imageability, namely ‘elephant’ and ‘tribune’. For example, the description of the item ‘elephant’ was ‘a large, grey animal with a trunk’. The mean concreteness rating was 3.97 (SD=0.96, range=1.53-4.93), the mean imageability rating of Van Loon-Vervoorn (1985) was 5.44 (SD=1.17, range=2.13-6.83)2. The
participants were unable to return to the previous description after they had entered a word. To control for any effects as a consequence of the sequence of the descriptions, the order of the descriptions was randomized for all participants.
2.2.3 Analyses
The answers of all the participants were checked both automatically and manually. If an answer exactly matched the expected answer it was counted as correct. For incorrect items a manual check was performed. When the participant made a typo, but it was clear that the participant meant the target word, the answer was considered correct. Plurals and diminutives were also considered correct. Sometimes the participants gave the verb instead of the noun or an appropriate synonym. These answers were also considered correct, since these words were about the same semantic concept as the target word. When a participant gave two words as an answer, this was considered false unless both words were correct answers. The two example descriptions are not included in the analysis. The responses from the participants are classified in three categories: target word, appropriate synonym and incorrect response / no response, see Appendix D. The inclusion criterion for an item was that at least 75% of the participants answered the description with the target word or an appropriate synonym. Items with less than 75% accuracy were disregarded in the further analyses.
It was decided to take the start of typing as the relevant reaction time rather than the time it took the participants to finish typing and submit the response. Both measures were highly correlated, r(1927) = .73, p < .001. The start reaction time is equal to how long the participants have thought about the semantic concept in their mind. By looking at the start reaction time, typing speed and word length are not influencing the performance. The start reaction time will from now on be referred to as reaction time. All reaction times below one second are excluded, since this was considered too short to read the description and retrieve a word to type. In these cases, the participant has probably accidentally started typing. Then the quartiles, the interquartile range (IR) between the first and third quartile and the strong outliers were calculated. After that, the strong outliers were excluded and the mean and SD for the reaction times and the three variables were calculated. Thereafter, the imageability ratings of Van Loon-Vervoorn (1985), the concreteness ratings of Brysbaert, et al. (2014) and the newly acquired imageability ratings have been converted to z-scores with the help of SPSS. The reason for this is that all three values had different scales; Van Loon-Vervoorn (1985) used a seven-point scale, Brysbaert, et al. (2014) used a five-seven-point scale and for the new imageability ratings we also used a five-point scale. After that, a stepwise multiple regression analysis has been conducted in SPSS to determine the relationship between the concreteness ratings of Brysbaert et al. (2014), the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings. A stepwise multiple regression analysis has been chosen in order to select the best fitting model and to control for potential multicollinearity. In SPSS the VIF values were determined and the VIF value for
2 Note. These values are based on 49 of the fifty items, because there was no imageability rating available for
the newly acquired imageability ratings turned out to be five, which means the coefficients and the p-values are questionable. Significance was met at 𝝰-level 0.05.
3. Results
First, the results of the questionnaire will be elaborated and thereafter the results of the participants on the naming task. Appendix A contains the newly acquired imageability ratings and Appendix D contains the responses of the participants on the naming task.
The qualitative analysis shows that eight items did not meet the inclusion criterion of 75% accurate response (target word or synonym). Table 1 shows these eight items along with the percentages.
Table 1.
Items that did not meet the inclusion criterion. Target
word Percentage correct
responses Uitstel 59% Kliekje 49% IJzel 71% Zintuig 51% Metselaar 65% Dooi 35% Luxe 73% Roepnaam 65%
3.1 New imageability ratings
3.1.1 Descriptive statistics questionnaire
The mean newly acquired imageability rating was 2,75 (SD=1.10, range=1.12-4.86). Appendix A contains the newly acquired imageability ratings per item.
3.1.2 Correlation analysis
To determine the correlation between the concreteness ratings of Brysbaert et al. (2014), the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings, a correlation analysis has been conducted. A significant strong positive correlation was found between the concreteness ratings of Brysbaert et al. (2014) and the imageability ratings of Van Loon-Vervoorn (1985), r(42) = .76, p < .001. The scatterplot below (figure 3) shows this positive correlation.
A significant, strong negative correlation was found between the concreteness ratings of Brysbaert et al. (2014) and the newly acquired imageability ratings, r (42) =- .77, p < .001. The scatterplot below (figure 4) shows this negative correlation.
Figure 4. Scatterplot concreteness Brysbaert et al. (2014) and new imageability ratings.
Finally, a significant, strong negative correlation was found between the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings, r(42) = -.83, p < .001. The scatterplot below (figure 5) shows this negative correlation.
Figure 5. Imageability Van Loon-Vervoorn (1985) and new imageability ratings
3.2 Naming task
3.2.1 Descriptive statistics naming task
The following outlier analysis for the reaction times is conducted. First, all reaction times below one second are excluded, since the participant has probably accidentally started typing. This concerned one item. Then the quartiles, the interquartile range (IR) and the strong outliers are calculated: Q1 = 2.27, Q3, = 4.30, IR = 2.03, strong outliers = (3 * 2.03) + 4.29) = 10.38 (upper bound) and 2.27 – (3 * 2.03) = -3.82 (lower bound). All reaction times above 10.38 are seen as strong outliers and are excluded. This concerned 79 items. No measurements were excluded at the lower bound. Table 2 shows the descriptive statistics of the reaction times and the three psycholinguistic variables.
Table 2.
Descriptive statistics
Min. Max. M SD
Reaction times 1.15 10.36 3.41 1.71
Concreteness 1.53 4.93 4.09 0.93
Imageability (Van Loon-Vervoorn, 1985) 2.33 6.83 5.68 1.00
New imageability ratings 1.12 4.86 2.50 0.98
3.2.2 Regression analysis
To determine the relationship between the reaction times and the concreteness ratings of Brysbaert et al. (2014), the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings, a stepwise multiple regression analysis has been conducted. The dependent variable was the reaction times and the independent variables were the three psycholinguistic variables. The stepwise linear regression showed that Concreteness and the new imageability ratings explain a significant amount of the variance of the reaction times (F(2, 1924) = 20.69, p < .001). The analysis shows that both Concreteness (Beta= 0.073; t (1924) = 4.431; p < .001) and the new imageability ratings (Beta = 0.082; t (1924) = 6.265; p < .001) significantly predict the reaction time. Imageability of Van Loon-Vervoorn (1985) does not meet the admission criterion and does not significantly predict the outcome (Beta = -0.025; t (1924) = -0.610; p = .542. The partial correlation between Imageability from Van Loon-Vervoorn (1985) and the reaction times is -.014, when corrected for the effect of Concreteness and the new imageability ratings. These results show that the Beta coefficient of the new imageability ratings is the highest, which means this variable explains the most variety.
The correlation analysis showed a very weak negative correlation between Concreteness and the reaction times, r(1927) = -.03, p = .074. This correlation is not significant. The scatterplot below (figure 6) shows this negative correlation.
Figure 6. Scatterplot reaction times and concreteness Brysbaert et al. (2014).
There was found a significant, very weak negative correlation between the imageability ratings of Van Loon-Vervoorn (1985) and the reaction times, r(1927) = -.09, p < .001. The scatterplot below (figure 7) shows this negative correlation.
Figure 7. Scatterplot reaction times and imageability Van Loon-Vervoorn (1985).
A significant, very weak positive correlation was found between the newly acquired imageability ratings and the reaction times, r(1927) = .11, p < .001. The scatterplot below (figure 8) shows this positive correlation.
Figure 8. Scatterplot reaction times and the new collected imageability ratings.
4. Discussion
The current study investigated the effect of Imageability and Concreteness on word retrieval in healthy individuals. A written naming task has been administered and the predictive value of the imageability ratings and the concreteness ratings has been evaluated. The naming task consisted of descriptions and the participants were asked to fill in the word matching the description. Furthermore, new imageability ratings for the used items have been acquired. A regression analysis and a correlation analysis have been conducted. The research question and subquestions were:
What is the influence of the psycholinguistic variables Imageability and Concreteness on word retrieval for healthy individuals?
- How strong is the correlation between Imageability and Concreteness?
- Are the imageability ratings of Van Loon-Vervoorn (1985) still valid or are they due for replacement?
- Does Concreteness or either of the two imageability ratings (Van Loon-Vervoorn (1985) vs. newly acquired ratings) have a higher predictive value for the reaction times in a word retrieval task than the other measures?
First, the coherence of the psycholinguistic variables will be discussed. Thereafter will be discussed if the imageability ratings of Van Loon-Vervoorn (1985) are due for replacement. Then will be discussed which of the three psycholinguistic variables is the better predictor and finally, limitations of the current study and recommendations for future research will be discussed.
4.1 Coherence variables
The correlation analysis shows that there is a significant strong positive correlation of .76 between the concreteness ratings of Brysbaert et al. (2014) and the imageability ratings of Van Loon-Vervoorn (1985). This means that a higher concreteness rating is associated with a higher imageability rating and a lower concreteness rating with a lower imageability rating. The correlation coefficient shows that 57,76% of the variance of the concreteness ratings can be explained by the imageability ratings. This result is in line with the hypothesis, that the variables Imageability and Concreteness are highly correlated. This is also in line with the literature, although correlations in the literature ranged from .78 to .85 (Friendly et al., 1982; Gilhooly & Logie, 1980; Paivio, 2013; Paivio et al., 1968) and the correlation in the current study falls just outside this range. Between the imageability ratings of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings there is a significant strong negative correlation of -.83. This means that a higher imageability from Van Loon-Vervoorn (1985) is associated with a higher new imageability or vice versa. The negative correlation is due to the inverse rating system. The correlation coefficient shows that 69,4% of the variance of the imageability ratings of Van Loon-Vervoorn (1985) can be explained by the new imageability ratings. This result is in line with the hypothesis, that the imageability ratings of Van Loon-Vervoorn (1985) are highly correlated with the newly acquired imageability ratings. The finding that the correlation between these two variables is strong and not very strong may be due to the fact that the imageability ratings of Van Loon-Vervoorn (1985) are already 35 years old. Perhaps these ratings are somewhat outdated. In the next chapter, this will be discussed in more detail. Between the concreteness ratings of Brysbaert et al. (2014) and the newly acquired imageability ratings there is also an significant strong negative correlation of -.77. This means that higher Concreteness is associated with higher Imageability. Again, the correlation is negative due to the inverse rating scale. The correlation coefficient shows that 60% of the variance of the concreteness ratings can be explained by the new imageability ratings. This result is in line with the hypothesis, that the variables Imageability and Concreteness are highly correlated. This is also in line with the literature, although correlations in the literature ranged from .78 to .85 (Friendly et al., 1982; Gilhooly & Logie, 1980; Paivio, 2013; Paivio et al., 1968) and this correlation again falls just outside this range. The correlation between Concreteness and the newly acquired imageability ratings is higher than the correlation between the concreteness ratings of Brysbaert et al. (2014) and the imageability ratings of Van Loon-Vervoorn (1985). This could be due to the fact that the participant groups of Brysbaert et al. (2014) and the current study are more equal. The participant group of Brysbaert et al. (2014) consisted of 74 participants (eleven men and 54 women) and our participant group consisted of 80 participants (fourteen men and 66 women). The participant group of Van Loon-Vervoorn (1985) consisted only of thirty participants and the gender distribution is very equal distributed (fifteen men and fifteen women). Another explanation is that the new imageability ratings are more up to date than the imageability ratings of Van Loon-Vervoorn (1985), and therefore have a stronger correlation with the concreteness ratings of Brysbaert et al. (2014).
Overall, the high correlations between Imageability and Concreteness in the current study are in line with the hypothesis and the literature (Friendly et al., 1982; Gilhooly & Logie, 1980; Paivio, 2013; Paivio et al., 1968). In general, words with a low concreteness also have a low imageability and words with a high concreteness have a high imageability, which also explains the high correlations. However, 30,6% - 42,2% of the variance in the imageability ratings cannot be explained by the concreteness ratings. This means that these two variables are not entirely the same, which is in line with the literature (Dellantonio et al., 2014; Paivio et al., 1968).
4.2 New imageability vs. old imageability (Van Loon-Vervoorn, 1985)
When comparing the newly acquired imageability ratings and the imageability ratings of Van Loon-Vervoorn (1985), it seems that the imageability ratings of Van Loon-Loon-Vervoorn (1985) are somewhat outdated. As described earlier, the correlation between these two variables is strong and not very strong, which could be due to the fact that the imageability ratings of Van Loon-Vervoorn (1985) are already 35 years old. Since Imageability gives an indication of how well a word gives rise to a mental image (Carroll & White, 1973), is it possible to believe that this mental image changes over the years. With the rise of the internet and social media during the past years, there are definitely words that nowadays evoke a clearer mental image than 35 years ago. Besides the fact that the imageability ratings of Van Loon-Vervoorn (1985) are 35 years old, the newly acquired imageability ratings were a better predictor for the reaction times than the imageability ratings of Van Loon-Vervoorn (1985). In the next chapter, this will be discussed in more detail. There are also differences in the way the ratings were collected. The participant group of Van Loon-Vervoorn (1985) consisted of thirty participants (fifteen men and fifteen women) that were all students Psychology or staff members from the psychological laboratory. Our participant group on the other hand consisted of eighty participants (fourteen men and 66 women) that were recruited in different areas. However, our participant group has a gender bias, since there were more female participants than male participants. The participant group of Van Loon-Vervoorn (1985) does not have this gender bias, but is smaller and more homogeneous than our participant group. Another difference between the current study and the study of Van Loon-Vervoorn (1985) is that in the current study a 5-point Likert scale has been used to acquire the imageability ratings and Van Loon-Vervoorn (1985) used a 7-point Likert scale. It is recommended to use a 5-point Likert scale, since this appears to be less confusing for the participants and increases response rate and response quality (Babakus & Mangold, 1992; Devlin et al., 2003; Hayes, 1992). Furthermore, the database of Van Loon-Vervoorn (1985) is not available online. This means each imageability rating has to be looked up manually, which makes it a very time-consuming process when one needs to look up multiple ratings. For these reasons, it is recommended to renew and expand the imageability ratings of Van Loon-Vervoorn (1985) and make them also available online.
4.3 Imageability vs. Concreteness
The stepwise multiple regression analysis showed that Concreteness and the new imageability ratings are both significant predictors of the reaction times of the participants on the naming task. The imageability ratings of Van Loon-Vervoorn (1985) did not meet the admission criterion for the analysis and appeared to be not a significant predictor of the reaction times of the participants on the naming task when the newly acquired ratings were already included. When looking at the Beta coefficient it is demonstrated that the new imageability ratings have more influence on the reaction times than Concreteness. This is not in line with the hypothesis that Concreteness can best explain the reaction times. This is also not in line with the literature, where there is more evidence for Concreteness to be the better predictor for reaction times (Guasch et al., 2016; Yao et al., 2017). When looking at the study of Guasch et al. (2016) and the study of Yao et al. (2017), there are a few differences with our study. First, both studies used another task to study the influence of Concreteness and Imageability on reaction times, namely a lexical decision task. When looking at the model of Ellis and Young (1988), a lexical decision task measures at an input level; a participant has only to decide whether a word exists or not. No semantic knowledge is necessary to fulfill the task, although it is about semantic variables that are being investigated. This is in contrast to a naming task, which measures at a semantic level and at an output level (Bastiaanse, 1991). It is assumed that the Semantic System also influences the Input Lexicon (the arrow in the model of Ellis and Young (1988) goes in two directions), but this has certainly less direct effect and may only take place if the support is needed to be able to do the task. In addition, there was a considerable amount of multicollinearity among the psycholinguistic variables in the study of Guasch et al. (2016) and they did not control for multicollinearity in their statistical analysis. This could have influenced the interpretation of the roles of the individual variables Concreteness and Imageability in the model.
The correlation analysis between the reaction times and Concreteness shows that there is a very weak negative correlation of -.03, which is not significant. When looking at the scatterplot (figure 6), it is visible that the reaction times tend to become shorter as a word becomes more concrete. However, a firm conclusion cannot be drawn since this correlation is not significant. Therefore, the hypothesis that healthy individuals have faster reaction times for concrete words cannot be confirmed. The correlation analysis between the reaction times and the imageability ratings of Van Loon-Vervoorn (1985) shows a significant very weak negative correlation of -.09. This means that the reaction times become shorter as the imageability of a word increases. This result is in line with the hypothesis and the literature, that healthy individuals have a faster reaction time for words with a high imageability (Strain et al., 1995). The correlation analysis between the reaction times and the newly acquired imageability ratings shows a significant very weak positive correlation of .11. This means that the reaction time becomes faster as the imageability of a word increases. This result is also in line with the hypothesis and the literature, that healthy individuals have a faster reaction time for words with a high imageability (Strain et al., 1995). That we found a positive correlation for the newly acquired imageability ratings and a negative correlation for the imageability ratings of Van Loon-Vervoorn (1985), is due to the fact that the rating scales had opposite directions. For the newly acquired imageability ratings, 1 was high imageable and 5 was low imageable, while for the imageability ratings of Van Loon-Vervoorn (1985) 1 was low imageable and 7 was high imageable. When comparing the correlations of Van Loon-Vervoorn (1985) and the newly acquired imageability ratings, it becomes clear that the newly acquired imageability ratings have clearly more influence on the reaction times than the imageability ratings of Van Loon-Vervoorn (1985). However, it should be taken into account that the correlations between the variables and the reaction times are very weak. There may have been other factors which have influenced the results, for example other psycholinguistic variables (i.e. Frequency or Age of Acquisition), typing speed of the participants, energy level of the participants during the experiment or the length of the descriptions of the items, causing the participants to lose focus.
These results show that Imageability is the better psycholinguistic variable to predict the reaction times on a written naming task. According to Dellantonio, Mulatti, Pastore and Job (2014), Imageability is more interesting than Concreteness, since Imageability does reflect the ease with which words evoke both body-external sensations and body-internal sensations. Therefore, for future research it is better to use Imageability than Concreteness to predict the outcomes for a word retrieval task.
4.4 Limitations and future research
A limitation of the current study is that it was an online study. Due to this, there was only limited control over the course of the study. Three other limitations were that the most participants in both participant groups were women, in the age range of 18-29 years old and had a high level of education, which makes the participant groups less heterogenous. A fourth limitation is that the native language of the participants on the naming task is not entirely certain. For future research it is recommended to ask the participants for their native language to ensure they really are native speakers. The last limitation of the current study is that some descriptions were not clear enough, requiring some items to be excluded for the analyses. The first recommendation for a follow up on the current study is to investigate the influence of Imageability and Concreteness on a verbal naming task in healthy individuals. Another recommendation for a follow up is to investigate the influence of Imageability and Concreteness on word retrieval in participants with language disorders, for example aphasia or dementia. It is already known in the literature that these two psycholinguistic variables affect naming performance in people with aphasia (Bastiaanse et al., 2016; Nickels & Howard, 1995), but it is still unknown which one is the better predictor for this population. A third recommendation is to investigate the influence of Imageability and Concreteness on word retrieval and investigate specifically the differences between nouns and verbs, since there are differences in the retrieval of verbs and nouns (Bastiaanse et al., 2016). It would also be interesting to investigate if there are any
differences in the variables Imageability and Concreteness on the performance of younger and older participants on a word retrieval task. When investigating the differences between the performance of younger and older participants, it can be seen whether aging affects the processing of these two variables. With the data of the current study it was presumably not possible to find age differences, since most of the participants had an age between 18-29 years and therefore the age distribution was not well distributed. Also, the influence of Imageability and Concreteness could be investigated on different tasks other than word retrieval tasks, for example synonym decision tasks or memory tasks to assess the influence in different modalities. The last recommendation is to update the imageability ratings of Van Loon-Vervoorn (1985). This is a rather old and small database, which is not available online. The newly acquired imageability ratings have led to a better model and therefore are the better predictor for the reaction times than the imageability ratings of Van Loon-Vervoorn (1985). Therefore, it is recommended to renew and expand the imageability ratings of Van Loon-Vervoorn (1985) and make them also available online. A disadvantage of renewing and expanding these imageability ratings is that it is very time consuming, since the participants have to rate a lot of words (Bysbaert et al., 2014). Finally, replicating the current study in other languages could be an interesting addition to the current evidence, to see if there are any differences between languages.
5. Conclusion
Overall, there can be concluded that Imageability is the better psycholinguistic variable to predict the reaction times of healthy individuals on a written naming task, specifically the newly acquired imageability ratings in the current study. From this it can be concluded that Imageability is the more predictive variable for word retrieval. Imageability and Concreteness are two strongly correlated psycholinguistic variables, but they are not the same and should definitely not be used interchangeably. It seems that it depends on the task which of the two variables best predicts the outcome. Furthermore, the imageability ratings of Van Loon-Vervoorn (1985) seem to be somewhat outdated and it is recommended to renew and expand this database. The current study adds new insights to the existing literature about the differences between the psycholinguistic variables Imageability and Concreteness in word retrieval, since previous research mostly focused on one of the two variables or the comparison of them in receptive rather than productive tasks.
References
Alario, F. X., Ferrand, L., Laganaro, M., New, B., Frauenfelder, U. H., & Segui, J. (2004). Predictors of picture naming speed. Behavior Research Methods, Instruments, and Computers, 36, 140–155. doi:10.3758/BF03195559
Babakus, E., & Mangold, W. G. (1992). Adapting the SERVQUAL scale to hospital services: an empirical investigation. Health Services Research, 26, 767–786.
Balota, D. A., Yap, M. J., Cortese, M. J., Hutchison, K. A., Kessler, B., Loftis, B., … Treiman, R. (2007). The english lexicon project. Behavior Research Methods, 39, 445–459. doi:10.3758/BF03193014 Barry, C., & Gerhand, S. (2003). Both concreteness and age-of-acquisition affect reading accuracy but
only concreteness affects comprehension in a deep dyslexic patient. Brain and Language, 84, 84–104. doi:10.1016/S0093-934X(02)00522-9
Bastiaanse, R. (1991). Afasie en taalverwerkingsmodellen. Tabu, 21, 127–140.
Bastiaanse, R. (2011). Afasie. Afasie. Houten: Bohn Stafleu van Loghum. doi:10.1007/978-90-313-9030-4
Bastiaanse, R., Bosje, M., & Visch-Brink, E. (1995). PALPA, vertaling in, aanpassing aan en normering voor het Nederlands van de PALPA, Psycholinguistic Assesments of Language Processing in Aphasia. Hove: Lawrence Erlbaum Associates.
Bastiaanse, R., Wieling, M., & Wolthuis, N. (2016). The role of frequency in the retrieval of nouns and verbs in aphasia. Aphasiology, 30, 1221–1239. doi:10.1080/02687038.2015.1100709
Biran, M., & Friedmann, N. (2005). From phonological paraphasias to the structure of the phonological output lexicon. Language and Cognitive Processes, 20, 589–616. doi:10.1080/01690960400005813
Bonin, P., Barry, C., Méot, A., & Chalard, M. (2004). The influence of age acquisition in word reading and other tasks: A never ending story? Journal of Memory and Language, 50, 456–476. doi:10.1016/j.jml.2004.02.001
Bonin, P., Méot, A., & Bugaiska, A. (2018). Concreteness norms for 1,659 French words: Relationships with other psycholinguistic variables and word recognition times. Behavior Research Methods, 50, 2366–2387. doi:10.3758/s13428-018-1014-y
Boyle, M. (2010). Semantic feature analysis treatment for aphasic word retrieval impairments: What’s in a name? Topics in Stroke Rehabilitation, 17, 411–422. doi:10.1310/tsr1706-411 Brunsdon, R., Coltheart, M., & Nickels, L. (2005). Treatment of irregular word spelling in
developmental surface dysgraphia. Cognitive Neuropsychology, 22, 213–251. doi:10.1080/02643290442000077
Brysbaert, M., & New, B. (2009). Moving beyond Kučera and Francis: A critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behavior Research Methods, 41, 977–990. doi:10.3758/BRM.41.4.977 Brysbaert, M., Van Wijnendaele, I., & De Deyne, S. (2000). Age-of-acquisition effects in semantic
processing tasks. Acta Psychologica, 104, 215–226. doi:10.1016/S0001-6918(00)00021-4 Brysbaert, M., Warriner, A. B., & Kuperman, V. (2014). Concreteness ratings for 40 thousand
generally known English word lemmas. Behavior Research Methods, 46, 904–911. doi:10.3758/s13428-013-0403-5
Butterworth, B., Howard, D., & Mcloughlin, P. (1984). The semantic deficit in aphasia: The relationship between semantic errors in auditory comprehension and picture naming. Neuropsychologia, 22, 409–426. doi:10.1016/0028-3932(84)90036-8
Bysbaert, M., Stevens, M., De Deyne, S., Voorspoels, W., & Storms, G. (2014). Norms of age of acquisition and concreteness for 30,000 Dutch words. Acta Psychologica, 150, 80–84. doi:10.1016/j.actpsy.2014.04.010
Carroll, J. B., & White, M. N. (1973). Word Frequency and Age of Acquisition as Determiners of Picture-Naming Latency. Quarterly Journal of Experimental Psychology, 25, 85–95. doi:10.1080/14640747308400325
CBS.
Chedid, G., Brambati, S. M., Bedetti, C., Rey, A. E., Wilson, M. A., & Vallet, G. T. (2019). Visual and auditory perceptual strength norms for 3,596 French nouns and their relationship with other psycholinguistic variables. Behavior Research Methods, 51, 2094–2105. doi:10.3758/s13428-019-01254-w
Code, C., & Herrmann, M. (2003). The relevance of emotional and psychosocial factors in aphasia to rehabilitation. Neuropsychological Rehabilitation, 13, 109–132.
doi:10.1080/09602010244000291
Connell, L., & Lynott, D. (2012). Strength of perceptual experience predicts word processing performance better than concreteness or imageability. Cognition, 125, 452–465. doi:10.1016/j.cognition.2012.07.010
Crutch, S. J., & Warrington, E. K. (2005). Abstract and concrete concepts have structurally different representational frameworks. Brain, 128, 615–627. doi:10.1093/brain/awh349
Dellantonio, S., Mulatti, C., Pastore, L., & Job, R. (2014). Measuring inconsistencies can lead you forward: Imageability and the X-ception theory. Frontiers in Psychology, 5, 1–9.
doi:10.3389/fpsyg.2014.00708
Devlin, S. J., Dong, H. K., & Brown, M. (2003). Selecting a Scale for Measuring Quality. Marketing Research, 5, 12–17.
Dorze, G. Le, & Nespoulous, J. L. (1989). Anomia in moderate aphasia: Problems in accessing the lexical representation. Brain and Language, 37, 381–400. doi:10.1016/0093-934X(89)90026-6 Ellis, A.W., Young, A. W. (1988). Human cognitive neuropsychology. Hove: Psychology Press. Ellis, A. W., & Young, A. W. (2014). Human cognitive neuropsychology, classic edition. Hove:
Psychology Press. doi:10.4324/9780203796382
Engelter, S. T., Gostynski, M., Papa, S., Frei, M., Born, C., Ajdacic-Gross, V., … Lyrer, P. A. (2006). Epidemiology of aphasia attributable to first ischemic stroke: Incidence, severity, fluency, etiology, and thrombolysis. Stroke, 37, 1379–1384. doi:10.1161/01.STR.0000221815.64093.8c Feiken, J., Hüttmann, A., & Links, P. (2020). Module specifieke therapie bij afasie. Assen: Uitgeverij
Koninklijke van Gorcum.
Ferrand, L., Brysbaert, M., Keuleers, E., New, B., Bonin, P., Méot, A., … Pallier, C. (2011). Comparing word processing times in naming, lexical decision, and progressive demasking: Evidence from Chronolex. Frontiers in Psychology, 2, 1–10. doi:10.3389/fpsyg.2011.00306
Ferrand, L., New, B., Brysbaert, M., Keuleers, E., Bonin, P., Méot, A., … Pallier, C. (2010). The French lexicon project: Lexical decision data for 38,840 French words and 38,840 pseudo words. Behavior Research Methods, 42, 488. doi:10.3758/BRM.42.2.488
Friendly, M., Franklin, P. E., Hoffman, D., & Rubin, D. C. (1982). The Toronto Word Pool: Norms for imagery, concreteness, orthographic variables, and grammatical usage for 1,080 words. Behavior Research Methods & Instrumentation, 14, 375–399. doi:10.3758/BF03203275
Gernsbacher, M. A. (1984). Resolving 20 years of inconsistent interactions between lexical familiarity and orthography, concreteness, and polysemy. Journal of Experimental Psychology: General, 113, 256–281. doi:10.1037/0096-3445.113.2.256
Gilhooly, K. J., & Logie, R. H. (1980). Age-of-acquisition, imagery, concreteness, familiarity, and ambiguity measures for 1,944 words. Behavior Research Methods & Instrumentation, 12, 95– 427. doi:10.3758/BF03201693
Guasch, M., Ferré, P., & Fraga, I. (2016). Spanish norms for affective and lexico-semantic variables for 1,400 words. Behavior Research Methods, 48, 358–1369. doi:10.3758/s13428-015-0684-y Hart, J., Berndt, R. S., & Caramazza, A. (1985). Category-specific naming deficit following cerebral
infarction. Nature, 316, 439–440. doi:10.1038/316439a0
Hashimoto, N., Widman, B., Kiran, S., & Richards, M. A. (2013). A comparison of features and categorical cues to improve naming abilities in aphasia. Aphasiology, 27, 1252–1279. doi:10.1080/02687038.2013.814760
Hayes, B. E. (1992). Measurement Customer Satisfaction: Development and Use of Questionnaire. Milwaukee, WI: ASQC Quality Press.
