Perception, 2013, volume 42, pages 1063 – 1074
doi:10.1068/p7525
The effects of perceptual priming on 4-year-olds’ haptic-to-visual cross-modal transfer Hilary Kalagher
Department of Psychology, Drew University, Madison, NJ 07940, USA; e‑mail: [email protected] Received 21 May 2013, in revised form 18 September 2013 Abstract. Four-year-old children often have difficulty visually recognizing objects that were previously experienced only haptically. This experiment attempts to improve their performance in these hapticto-visual transfer tasks. Sixty-two 4-year-old children participated in priming trials in which they explored eight unfamiliar objects visually, haptically, or visually and haptically together. Subsequently, all children participated in the same haptic-to-visual cross-modal transfer task. In this task, children haptically explored the objects that were presented in the priming phase and then visually identified a match from among three test objects, each matching the object on only one dimension (shape, texture, or color). Children in all priming conditions predominantly made shape-based matches; however, the most shape-based matches were made in the Visual and Haptic condition. All kinds of priming provided the necessary memory traces upon which subsequent haptic exploration could build a strong enough representation to enable subsequent visual recognition. Haptic exploration patterns during the cross-modal transfer task are discussed and the detailed analyses provide a unique contribution to our understanding of the development of haptic exploratory procedures. Keywords: cross-modal transfer, haptic exploration, perceptual priming, children
1 Introduction Perceptual priming is the phenomenon in which the perception or identification of an object is improved as a result of exposure to, or experience with, that object. The effects of perceptual priming are long lasting and robust (Cave 1997). The development of perceptual priming appears to be relatively stable from middle childhood through late adulthood (Ballesteros and Mayas 2009; Ballesteros et al 2007; Billingsley et al 2002; Russo et al 1995; Wippich 1991). Moreover, there is evidence to suggest that perceptual priming is intact in the early stages of Alzheimer’s disease (Ballesteros and Reales 2004) and in amnesic patients (Cave and Squire 1992; Hamann and Squire 1997). Relatively few studies, however, have looked at perceptual priming in children younger than middle childhood (Hayes and Hennessy 1996; Parkin and Streete 1988; Rovee-Collier 1997). Parkin and Streete (1988) provided a notable exception in which 3-year-olds, 5-yearolds, and 7-year-olds, and adults participated in a picture completion task. The learning phase consisted of pictorial perceptual priming in which participants were visually presented with a series of pictures in their most incomplete form. The pictures were then visually presented in increasingly detailed forms until identification was made. The same series of pictures were visually presented at a later time (after either 1 h or 2 weeks). Identification of a less detailed picture in the test phase than during the learning phase was taken as evidence of perceptual priming. All age groups demonstrated perceptual priming; however, the performance of 3-year-olds was poorer. Similarly, Hayes and Hennessy (1996) measured picture identification thresholds for 4-, 5-, and 10-year-old children in an experimental setup similar to Parkin and Streete (1988) with a delay period of 48 h. Varying levels of perceptual similarity between studied items and tested items were used. Each of the studied and tested item pairs were from the same basic level category, but half of them shared high perceptual similarity whereas the other half shared
1064
H Kalagher
low perceptual similarity. Both studied and tested items were presented visually. Hayes and Hennessy (1996) reported greater priming effects for the category of test items that shared the most perceptual similarity with the learned items. These authors concluded that priming is a function of storage and generalization of specific perceptual features and not by abstract representations. Taken together, these studies provide clear evidence that perceptual priming is present in children as young as 4-years-old; however, it is sensitive to the specific perceptual features encountered during study. The studies just reported examined the effects of visual perceptual priming on a visual recognition task. Less is known about the extent to which information presented in one modality could facilitate recognition in another modality in children. This question has been studied extensively in adults (eg Ballesteros 2008; Craddock and Lawson 2008; Easton et al 1997; Reales and Ballesteros 1999), where there is converging evidence that cross-modal priming is as effective as intramodal priming for enhancing object recognition. Moreover, the results of cross-modal facilitation in object recognition between haptics and vision are consistent with evidence of successful cross-modal transfer in adults (eg Gaydos 1956; Kalagher and Jones 2011a; Norman et al 2006). The status of cross-modal transfer in childhood, however, is unclear. A body of evidence now exists on the successful cross-modal transfer of information in prelinguistic infants (eg Streri and Gentaz 2003). However, this finding is not robust and there appear to be a large amount of interindividual differences that are present early in life (cf Rose 1994). Preschool-aged children often have difficulty visually identifying objects that were initially explored haptically. Attempts to explain these difficulties largely center around two, not mutually exclusive, explanations. Some have suggested that this difficulty exists because children form qualitatively different representations from visual experiences and haptic experiences. For example, information obtained haptically might largely consist of texture information and therefore might not be easily used in a visual task because in vision we rely more on shape information (eg Abravanel 1968; Bushnell and Baxt 1999). Others have suggested that deficits in haptic-to-visual transfer are the result of poor haptic abilities (eg Kalagher and Jones 2011a; Milner and Bryant 1970; Rose et al 1972; Scofield et al 2009). Poor haptic abilities could be the result of a difficulty in forming percepts from haptically obtained information; or they might be because children do not execute mature haptic exploration behavior which results in incomplete or inadequate information. The goal of the present experiment is to inform the discussion of these proposals by extending the work of Kalagher and Jones (2010, 2011a); therefore, their work will be reviewed in detail. Kalagher and Jones (2011a) asked children aged 2½, 3, 3½, 4, 4½, and 5 years and adults to complete a visual-to-visual intramodal task and a haptic-to-visual crossmodal task. A novel name extension paradigm was used in which children were introduced to eight novel category exemplars. On each of 8 trials, participants explored an exemplar object either haptically or visually and were told its novel name (eg “That is a dax”). After the exemplar object was removed, participants were presented visually with three test objects and were asked to indicate which object shared the same name as the exemplar (eg “Which one is the dax?”). Each of the 3‑D test objects shared a different attribute (shape, texture, or color) with its exemplar and differed from the exemplar and the other two test objects on the remaining two dimensions. If qualitatively different representations are formed through visual and haptic exploration, then children would make different kinds of matches following visual and haptic exploration. In the visual-to-visual task, shape-based matches were predominantely made by all age groups. However, only 5-year-olds and adults made shape-based matches in the haptic-to‑visual task; children younger than 5 years old did not make systematic matches— they were equally likely to choose any of the three test objects (Kalagher and Jones 2011a).
Priming effects
1065
Using the identical paradigm and stimuli, Kalagher and Jones (2010) tested 4‑year-old children in visual-to-haptic and haptic-to-haptic conditions of transfer. Interestingly, children in this experiment also made systematic matches for the category exemplars. Thus, it is only in the haptic-to-visual transfer task of this paradigm that 4-year-old children have not demonstrated systematic matching behavior. Importantly, Kalagher and Jones (2011a) also examined the kinds of exploration patterns children used while haptically exploring the objects. Analysis of children’s hand movements during haptic exploration found no instances of the hand actions in Lederman and Klatzky’s (1987) taxonomy of adult exploratory procedures but did find that certain hand movements produced by children were reliably associated with subsequent object matches. However, children younger than 5 years old produced these movements with very low frequency. Thus, Kalagher and Jones (2011a) concluded that the difficulties in haptic-to-visual transfer are the result of poor haptic perception, however, because systematic matches were not made, conclusions about the extent to which representations formed through visual or haptic exploration are qualitatively different could not be drawn. Because visual priming facilitates 4‑year-old children’s visual recognition of stimuli, in this study I ask if priming (visual, haptic, and visual and haptic) will increase the likelihood that 4-year-olds will make systematic matches in a haptic-to-visual task. The current experiment differs from that of Kalagher and Jones (2011a) in two ways. First, a priming phase in which children experienced one of three different kinds of priming—visual only, haptic only, and visual and haptic—was included. After priming, all children completed a haptic-to-visual matching task identical to that in Kalagher and Jones. The addition of cross-modal priming conditions is a unique contribution of the present experiment, as it has not been studied in 4‑year-old children. The second difference is that this study focuses on 4‑year-olds only. Kalagher and Jones (2010, 2011a) found systematic matching behavior in 4‑year-olds in their visual-to-visual, visual-to-haptic, and haptic-to-haptic versions of the novel name extension paradigm, but not in the haptic-to-visual version. It is also the age where there is converging evidence that intramodal visual priming is effective (Hayes and Hennessy 1996; Parkin and Streete 1988), suggesting that children at this age might benefit from cross-modal priming. The present experiment asks three questions: (i) Will priming affect performance in a subsequent haptic-to visual task resulting in a pattern of systematic matching behavior? (ii) Will patterns of matching behavior differ as a function of priming condition? (iii) Will children execute adult-like haptic exploratory procedures? 2 Method 2.1 Participants Sixty-two 4-year-olds (M = 52.6 months; range = 46.4–60.4 months; twenty-nine males) participated in this study. Although the range suggests that 5‑year-old children were also included, only one child exceeded 60 months of age. Participants reflected the local community in social class, ethnicity, and racial identity; nearly all participants were from White, middleclass families. Children were randomly assigned to one of three priming conditions (“Visual Only”, n = 21; “Haptic Only”, n = 18; or “Visual and Haptic”, n = 23). 2.2 Stimuli The stimulus set consisted of eight unfamiliar category exemplars, three test objects for each exemplar, and nine familiar objects (eg cup, spoon, comb). Exemplars were three-dimensional nonsense objects constructed from a variety of materials, including wood, clay, and cloth. Sizes ranged from 7 cm to 17 cm. Colors, textures, and shapes were widely varied. Each of the test objects shared a different attribute—shape, texture, or color—with its exemplar and differed from the exemplar and the other two test objects on the other two dimensions (see figure 1
1066
H Kalagher
for a sample stimulus set). Color, of course, cannot be perceived haptically. Previous studies of a similar nature, (eg Kalagher and Jones 2010, 2011a; Smith et al 2002) have provided children with three perceptual features on which to match (shape, color, and texture) so as to best distinguish a pattern of perceptually based object matches from chance performance.
Category exemplar
Test object 1: color match
Test object 2: shape match
Test object 3: texture match
Figure 1. Sample stimuli set: one exemplar and three test objects, each matching the exemplar on one dimension—shape, texture, or color—and differing from the exemplar and each other on the other two dimensions.
2.3 Procedure 2.3.1 Overview of design. Each child was seated at a table next to his or her parent and across from the experimenter. All children completed two blocks of trials: (i) the priming block (eight trials) and (ii) the testing block (3 training trials and 8 testing trials). The digital records of the experiment were coded for the study’s two dependent measures: (i) choice of object match (shape, texture, or color) and (ii) the kinds of haptic exploratory procedures produced and the frequency with which they occurred. The entire experiment took approximately 20 min to complete. Priming trials. The first block consisted of 8 priming trials; on each trial the children explored one object. The priming trials differed depending upon condition. In the Visual and Haptic condition, children were presented with each object one at a time and encouraged to explore it for approximately 10 s (M = 8.02 s; range = 1.7–31.5 s; SD = 4.25 s). In the Visual Only condition, children were handed each object, which was placed in a clear plastic sphere (see figure 2a for Visual Only stimulus set), one at time. [Objects were placed in a clear plastic sphere so that only visual information (and not haptic) could be obtained from exploration.] Children were encouraged to play with each sphere for approximately 10 s (M = 7.76 s; range = 2–16.3 s; SD = 2.4 s). In the Haptic Only condition, children placed their hands and forearms into a box, and a piece of cloth was pulled over the child’s arms to prevent him or her from seeing inside the box. Each object was placed one at a time into the hands of the child inside the box. Children were encouraged to explore each object for 10 s (M = 7.94 s; range = 3.3–18.1 s; SD = 2.42 s). Objects were presented in the same order across all conditions and were never named during priming. The objects used for priming in the Haptic Only and the Visual and Haptic conditions were the same (see figure 2b for the Haptic Only and Visual and Haptic stimulus sets) and were the exact objects used as category exemplars in the second block of trials.
(a)
(b) Figure 2. Stimuli used during familiarization phases for the (a) Visual Only condition and for the (b) Haptic Only and the Visual and Haptic conditions.
Priming effects
1067
Testing trials. The second block consisted of 3 training trials and 8 testing trials and was the same for all children regardless of priming condition. The testing trials were identical to those used in Kalagher and Jones (2011a), allowing for more accurate comparisons to be drawn between their results and those of the present study. A novel name extension paradigm was used in which children were introduced to novel object categories through a single category exemplar and asked to extend that novel name to one of three test objects. The testing phase began with three training trials to ensure that the child understood the task. In each training trial, the child placed his or her hands and forearms inside the box, and a piece of cloth was pulled down to prevent viewing. The experimenter put a familiar object into the hands of the child within the box, identified it by name, and asked the child if he or she could feel it (eg “This is a spoon. Can you feel the spoon?”). The object was removed from the box after 5 s and the child was visually presented with three test objects (eg a cup, a spoon, and a comb) and asked to indicate the test object with the same name as the exemplar object (eg “Can you show me the spoon?”). Children indicated their choice of a matching object by removing their hands from the box and using them to point. Test trials followed immediately and were structured in the same way as the training trials. The child was handed the exemplar object from one set at a time inside the box and told its novel name (eg “This is a teeka”). Children were given 5 s to explore the exemplar, after which the object was removed from the box. The three test items were then visually presented to the child; all items were within the child’s reach. The object category name was again used when the experimenter asked for a match (eg “Can you show me another teeka?”). 2.3.2 Behavioral coding. A small camera recorded the experiment. Coders recorded the two dependent measures: the match choice and the kinds and number of haptic exploratory procedures produced. Match choice. The first item touched, handed over, or pointed to was scored as the child’s choice of match. No child failed to indicate a choice by one of these behaviors on any trial. Children were given a sticker after each trial regardless of the choice they made. Coders indicated the test object—shape match, texture match, or color match—chosen on each trial. There were no ‘correct’ kinds of matches; each option was a correct match on a different dimension. However, color matches could not reflect the perception of color during exploration and so were interpreted as random choices. Hand exploratory procedures. The digital recordings were coded for children’s hand movements while they explored the objects in the cross-modal task. Digital recordings were uploaded into the Noldus Observer XT software (Noldus Information Technology Inc, Leesburg, VA), a behavioral coding system. The software allows the recording to be slowed down. Coders slowed the recordings down to 1/20 time and looked for instances of six of the eight exploratory procedures identified by Lederman and Klatzky (1987) in their adult sample: (i) contour following, associated with searches for exact shape information; (ii) lateral motion, associated with searches for texture information; (iii) pressure, associated with searches for information about rigidity; (iv) unsupported holding, associated with searches for weight information; (v) enclosure, associated with searches for global shape information; and (vi) static contact associated with searches for temperature information. The other two exploratory procedures were associated with object part motion and function and, therefore, were not appropriate for the stimulus objects. Coders were also instructed to look for movements that were not in Lederman and Klatzky’s original taxonomy. The only additional movement recorded more than once was pinching. This movement occurred at a very low frequency (M = 0.02) and was therefore not analyzed further. Coders considered exploratory procedures to be mutually exclusive. The hand-movement coding yielded the frequencies with which a participant produced each of the six exploratory procedures.
1068
H Kalagher
The primary coder coded all of the trials while the secondary coder randomly selected trials that were evenly distributed among the conditions to code for reliability. Hand movements were clearly visible on 429 trials (86%). The remaining 67 trials could not be coded because of experimental error (9 trials) or because the children’s hands were too far back from the camera’s angle to be clearly seen (58 trials). 96 (22%) of the 429 trials randomly selected trials showed satisfactory intercoder agreement on whether or not a specific exploratory procedure was present on each trial (Cohen’s l = 0.53). 3 Results 3.1 Priming conditions The exploration time in each of the three priming conditions was examined to ensure that priming phases were similar. To perform this analysis, the data were organized such that each familiarization trial was its own data point. This allowed for the comparison of exploration time both among conditions, and also among the eight different test stimuli. There were a total of 496 trials; 7 of those trials were thrown out due to experimenter error, leaving 489 trials to analyze. (The means, standard deviations, and ranges of exploration times for each of the three conditions were reported in section 2.) A 3 condition (Visual Only, Haptic Only, and Visual and Haptic) by 8 stimuli ANOVA was conducted with the time spent exploring as the dependent measure. There was no difference in exploration time among the three priming conditions, (F2, 487 = 0.309, ns) or among the eight different stimuli (F7, 482 = 0.409, ns), and no interaction. Because the time spent exploring the objects was similar in all three conditions and across the eight different stimuli, differences found in object matching can be attributed to the different experiences afforded by the priming conditions. 3.2 Object matching (i) Will priming affect performance in a subsequent haptic-to-visual task resulting in a pattern of systematic matching behavior? Figure 3 shows the means with which children chose shape (white bars), texture (striped bars), and color (dotted bars) matches for the exemplars in each of the three conditions. The dashed line shows the chance level of performance (2.67) given three choices. Onesample t‑tests confirm the impression from figure 3 that shape choices in each condition (Visual Only: t20 = 6.46, p
doi:10.1068/p7525
The effects of perceptual priming on 4-year-olds’ haptic-to-visual cross-modal transfer Hilary Kalagher
Department of Psychology, Drew University, Madison, NJ 07940, USA; e‑mail: [email protected] Received 21 May 2013, in revised form 18 September 2013 Abstract. Four-year-old children often have difficulty visually recognizing objects that were previously experienced only haptically. This experiment attempts to improve their performance in these hapticto-visual transfer tasks. Sixty-two 4-year-old children participated in priming trials in which they explored eight unfamiliar objects visually, haptically, or visually and haptically together. Subsequently, all children participated in the same haptic-to-visual cross-modal transfer task. In this task, children haptically explored the objects that were presented in the priming phase and then visually identified a match from among three test objects, each matching the object on only one dimension (shape, texture, or color). Children in all priming conditions predominantly made shape-based matches; however, the most shape-based matches were made in the Visual and Haptic condition. All kinds of priming provided the necessary memory traces upon which subsequent haptic exploration could build a strong enough representation to enable subsequent visual recognition. Haptic exploration patterns during the cross-modal transfer task are discussed and the detailed analyses provide a unique contribution to our understanding of the development of haptic exploratory procedures. Keywords: cross-modal transfer, haptic exploration, perceptual priming, children
1 Introduction Perceptual priming is the phenomenon in which the perception or identification of an object is improved as a result of exposure to, or experience with, that object. The effects of perceptual priming are long lasting and robust (Cave 1997). The development of perceptual priming appears to be relatively stable from middle childhood through late adulthood (Ballesteros and Mayas 2009; Ballesteros et al 2007; Billingsley et al 2002; Russo et al 1995; Wippich 1991). Moreover, there is evidence to suggest that perceptual priming is intact in the early stages of Alzheimer’s disease (Ballesteros and Reales 2004) and in amnesic patients (Cave and Squire 1992; Hamann and Squire 1997). Relatively few studies, however, have looked at perceptual priming in children younger than middle childhood (Hayes and Hennessy 1996; Parkin and Streete 1988; Rovee-Collier 1997). Parkin and Streete (1988) provided a notable exception in which 3-year-olds, 5-yearolds, and 7-year-olds, and adults participated in a picture completion task. The learning phase consisted of pictorial perceptual priming in which participants were visually presented with a series of pictures in their most incomplete form. The pictures were then visually presented in increasingly detailed forms until identification was made. The same series of pictures were visually presented at a later time (after either 1 h or 2 weeks). Identification of a less detailed picture in the test phase than during the learning phase was taken as evidence of perceptual priming. All age groups demonstrated perceptual priming; however, the performance of 3-year-olds was poorer. Similarly, Hayes and Hennessy (1996) measured picture identification thresholds for 4-, 5-, and 10-year-old children in an experimental setup similar to Parkin and Streete (1988) with a delay period of 48 h. Varying levels of perceptual similarity between studied items and tested items were used. Each of the studied and tested item pairs were from the same basic level category, but half of them shared high perceptual similarity whereas the other half shared
1064
H Kalagher
low perceptual similarity. Both studied and tested items were presented visually. Hayes and Hennessy (1996) reported greater priming effects for the category of test items that shared the most perceptual similarity with the learned items. These authors concluded that priming is a function of storage and generalization of specific perceptual features and not by abstract representations. Taken together, these studies provide clear evidence that perceptual priming is present in children as young as 4-years-old; however, it is sensitive to the specific perceptual features encountered during study. The studies just reported examined the effects of visual perceptual priming on a visual recognition task. Less is known about the extent to which information presented in one modality could facilitate recognition in another modality in children. This question has been studied extensively in adults (eg Ballesteros 2008; Craddock and Lawson 2008; Easton et al 1997; Reales and Ballesteros 1999), where there is converging evidence that cross-modal priming is as effective as intramodal priming for enhancing object recognition. Moreover, the results of cross-modal facilitation in object recognition between haptics and vision are consistent with evidence of successful cross-modal transfer in adults (eg Gaydos 1956; Kalagher and Jones 2011a; Norman et al 2006). The status of cross-modal transfer in childhood, however, is unclear. A body of evidence now exists on the successful cross-modal transfer of information in prelinguistic infants (eg Streri and Gentaz 2003). However, this finding is not robust and there appear to be a large amount of interindividual differences that are present early in life (cf Rose 1994). Preschool-aged children often have difficulty visually identifying objects that were initially explored haptically. Attempts to explain these difficulties largely center around two, not mutually exclusive, explanations. Some have suggested that this difficulty exists because children form qualitatively different representations from visual experiences and haptic experiences. For example, information obtained haptically might largely consist of texture information and therefore might not be easily used in a visual task because in vision we rely more on shape information (eg Abravanel 1968; Bushnell and Baxt 1999). Others have suggested that deficits in haptic-to-visual transfer are the result of poor haptic abilities (eg Kalagher and Jones 2011a; Milner and Bryant 1970; Rose et al 1972; Scofield et al 2009). Poor haptic abilities could be the result of a difficulty in forming percepts from haptically obtained information; or they might be because children do not execute mature haptic exploration behavior which results in incomplete or inadequate information. The goal of the present experiment is to inform the discussion of these proposals by extending the work of Kalagher and Jones (2010, 2011a); therefore, their work will be reviewed in detail. Kalagher and Jones (2011a) asked children aged 2½, 3, 3½, 4, 4½, and 5 years and adults to complete a visual-to-visual intramodal task and a haptic-to-visual crossmodal task. A novel name extension paradigm was used in which children were introduced to eight novel category exemplars. On each of 8 trials, participants explored an exemplar object either haptically or visually and were told its novel name (eg “That is a dax”). After the exemplar object was removed, participants were presented visually with three test objects and were asked to indicate which object shared the same name as the exemplar (eg “Which one is the dax?”). Each of the 3‑D test objects shared a different attribute (shape, texture, or color) with its exemplar and differed from the exemplar and the other two test objects on the remaining two dimensions. If qualitatively different representations are formed through visual and haptic exploration, then children would make different kinds of matches following visual and haptic exploration. In the visual-to-visual task, shape-based matches were predominantely made by all age groups. However, only 5-year-olds and adults made shape-based matches in the haptic-to‑visual task; children younger than 5 years old did not make systematic matches— they were equally likely to choose any of the three test objects (Kalagher and Jones 2011a).
Priming effects
1065
Using the identical paradigm and stimuli, Kalagher and Jones (2010) tested 4‑year-old children in visual-to-haptic and haptic-to-haptic conditions of transfer. Interestingly, children in this experiment also made systematic matches for the category exemplars. Thus, it is only in the haptic-to-visual transfer task of this paradigm that 4-year-old children have not demonstrated systematic matching behavior. Importantly, Kalagher and Jones (2011a) also examined the kinds of exploration patterns children used while haptically exploring the objects. Analysis of children’s hand movements during haptic exploration found no instances of the hand actions in Lederman and Klatzky’s (1987) taxonomy of adult exploratory procedures but did find that certain hand movements produced by children were reliably associated with subsequent object matches. However, children younger than 5 years old produced these movements with very low frequency. Thus, Kalagher and Jones (2011a) concluded that the difficulties in haptic-to-visual transfer are the result of poor haptic perception, however, because systematic matches were not made, conclusions about the extent to which representations formed through visual or haptic exploration are qualitatively different could not be drawn. Because visual priming facilitates 4‑year-old children’s visual recognition of stimuli, in this study I ask if priming (visual, haptic, and visual and haptic) will increase the likelihood that 4-year-olds will make systematic matches in a haptic-to-visual task. The current experiment differs from that of Kalagher and Jones (2011a) in two ways. First, a priming phase in which children experienced one of three different kinds of priming—visual only, haptic only, and visual and haptic—was included. After priming, all children completed a haptic-to-visual matching task identical to that in Kalagher and Jones. The addition of cross-modal priming conditions is a unique contribution of the present experiment, as it has not been studied in 4‑year-old children. The second difference is that this study focuses on 4‑year-olds only. Kalagher and Jones (2010, 2011a) found systematic matching behavior in 4‑year-olds in their visual-to-visual, visual-to-haptic, and haptic-to-haptic versions of the novel name extension paradigm, but not in the haptic-to-visual version. It is also the age where there is converging evidence that intramodal visual priming is effective (Hayes and Hennessy 1996; Parkin and Streete 1988), suggesting that children at this age might benefit from cross-modal priming. The present experiment asks three questions: (i) Will priming affect performance in a subsequent haptic-to visual task resulting in a pattern of systematic matching behavior? (ii) Will patterns of matching behavior differ as a function of priming condition? (iii) Will children execute adult-like haptic exploratory procedures? 2 Method 2.1 Participants Sixty-two 4-year-olds (M = 52.6 months; range = 46.4–60.4 months; twenty-nine males) participated in this study. Although the range suggests that 5‑year-old children were also included, only one child exceeded 60 months of age. Participants reflected the local community in social class, ethnicity, and racial identity; nearly all participants were from White, middleclass families. Children were randomly assigned to one of three priming conditions (“Visual Only”, n = 21; “Haptic Only”, n = 18; or “Visual and Haptic”, n = 23). 2.2 Stimuli The stimulus set consisted of eight unfamiliar category exemplars, three test objects for each exemplar, and nine familiar objects (eg cup, spoon, comb). Exemplars were three-dimensional nonsense objects constructed from a variety of materials, including wood, clay, and cloth. Sizes ranged from 7 cm to 17 cm. Colors, textures, and shapes were widely varied. Each of the test objects shared a different attribute—shape, texture, or color—with its exemplar and differed from the exemplar and the other two test objects on the other two dimensions (see figure 1
1066
H Kalagher
for a sample stimulus set). Color, of course, cannot be perceived haptically. Previous studies of a similar nature, (eg Kalagher and Jones 2010, 2011a; Smith et al 2002) have provided children with three perceptual features on which to match (shape, color, and texture) so as to best distinguish a pattern of perceptually based object matches from chance performance.
Category exemplar
Test object 1: color match
Test object 2: shape match
Test object 3: texture match
Figure 1. Sample stimuli set: one exemplar and three test objects, each matching the exemplar on one dimension—shape, texture, or color—and differing from the exemplar and each other on the other two dimensions.
2.3 Procedure 2.3.1 Overview of design. Each child was seated at a table next to his or her parent and across from the experimenter. All children completed two blocks of trials: (i) the priming block (eight trials) and (ii) the testing block (3 training trials and 8 testing trials). The digital records of the experiment were coded for the study’s two dependent measures: (i) choice of object match (shape, texture, or color) and (ii) the kinds of haptic exploratory procedures produced and the frequency with which they occurred. The entire experiment took approximately 20 min to complete. Priming trials. The first block consisted of 8 priming trials; on each trial the children explored one object. The priming trials differed depending upon condition. In the Visual and Haptic condition, children were presented with each object one at a time and encouraged to explore it for approximately 10 s (M = 8.02 s; range = 1.7–31.5 s; SD = 4.25 s). In the Visual Only condition, children were handed each object, which was placed in a clear plastic sphere (see figure 2a for Visual Only stimulus set), one at time. [Objects were placed in a clear plastic sphere so that only visual information (and not haptic) could be obtained from exploration.] Children were encouraged to play with each sphere for approximately 10 s (M = 7.76 s; range = 2–16.3 s; SD = 2.4 s). In the Haptic Only condition, children placed their hands and forearms into a box, and a piece of cloth was pulled over the child’s arms to prevent him or her from seeing inside the box. Each object was placed one at a time into the hands of the child inside the box. Children were encouraged to explore each object for 10 s (M = 7.94 s; range = 3.3–18.1 s; SD = 2.42 s). Objects were presented in the same order across all conditions and were never named during priming. The objects used for priming in the Haptic Only and the Visual and Haptic conditions were the same (see figure 2b for the Haptic Only and Visual and Haptic stimulus sets) and were the exact objects used as category exemplars in the second block of trials.
(a)
(b) Figure 2. Stimuli used during familiarization phases for the (a) Visual Only condition and for the (b) Haptic Only and the Visual and Haptic conditions.
Priming effects
1067
Testing trials. The second block consisted of 3 training trials and 8 testing trials and was the same for all children regardless of priming condition. The testing trials were identical to those used in Kalagher and Jones (2011a), allowing for more accurate comparisons to be drawn between their results and those of the present study. A novel name extension paradigm was used in which children were introduced to novel object categories through a single category exemplar and asked to extend that novel name to one of three test objects. The testing phase began with three training trials to ensure that the child understood the task. In each training trial, the child placed his or her hands and forearms inside the box, and a piece of cloth was pulled down to prevent viewing. The experimenter put a familiar object into the hands of the child within the box, identified it by name, and asked the child if he or she could feel it (eg “This is a spoon. Can you feel the spoon?”). The object was removed from the box after 5 s and the child was visually presented with three test objects (eg a cup, a spoon, and a comb) and asked to indicate the test object with the same name as the exemplar object (eg “Can you show me the spoon?”). Children indicated their choice of a matching object by removing their hands from the box and using them to point. Test trials followed immediately and were structured in the same way as the training trials. The child was handed the exemplar object from one set at a time inside the box and told its novel name (eg “This is a teeka”). Children were given 5 s to explore the exemplar, after which the object was removed from the box. The three test items were then visually presented to the child; all items were within the child’s reach. The object category name was again used when the experimenter asked for a match (eg “Can you show me another teeka?”). 2.3.2 Behavioral coding. A small camera recorded the experiment. Coders recorded the two dependent measures: the match choice and the kinds and number of haptic exploratory procedures produced. Match choice. The first item touched, handed over, or pointed to was scored as the child’s choice of match. No child failed to indicate a choice by one of these behaviors on any trial. Children were given a sticker after each trial regardless of the choice they made. Coders indicated the test object—shape match, texture match, or color match—chosen on each trial. There were no ‘correct’ kinds of matches; each option was a correct match on a different dimension. However, color matches could not reflect the perception of color during exploration and so were interpreted as random choices. Hand exploratory procedures. The digital recordings were coded for children’s hand movements while they explored the objects in the cross-modal task. Digital recordings were uploaded into the Noldus Observer XT software (Noldus Information Technology Inc, Leesburg, VA), a behavioral coding system. The software allows the recording to be slowed down. Coders slowed the recordings down to 1/20 time and looked for instances of six of the eight exploratory procedures identified by Lederman and Klatzky (1987) in their adult sample: (i) contour following, associated with searches for exact shape information; (ii) lateral motion, associated with searches for texture information; (iii) pressure, associated with searches for information about rigidity; (iv) unsupported holding, associated with searches for weight information; (v) enclosure, associated with searches for global shape information; and (vi) static contact associated with searches for temperature information. The other two exploratory procedures were associated with object part motion and function and, therefore, were not appropriate for the stimulus objects. Coders were also instructed to look for movements that were not in Lederman and Klatzky’s original taxonomy. The only additional movement recorded more than once was pinching. This movement occurred at a very low frequency (M = 0.02) and was therefore not analyzed further. Coders considered exploratory procedures to be mutually exclusive. The hand-movement coding yielded the frequencies with which a participant produced each of the six exploratory procedures.
1068
H Kalagher
The primary coder coded all of the trials while the secondary coder randomly selected trials that were evenly distributed among the conditions to code for reliability. Hand movements were clearly visible on 429 trials (86%). The remaining 67 trials could not be coded because of experimental error (9 trials) or because the children’s hands were too far back from the camera’s angle to be clearly seen (58 trials). 96 (22%) of the 429 trials randomly selected trials showed satisfactory intercoder agreement on whether or not a specific exploratory procedure was present on each trial (Cohen’s l = 0.53). 3 Results 3.1 Priming conditions The exploration time in each of the three priming conditions was examined to ensure that priming phases were similar. To perform this analysis, the data were organized such that each familiarization trial was its own data point. This allowed for the comparison of exploration time both among conditions, and also among the eight different test stimuli. There were a total of 496 trials; 7 of those trials were thrown out due to experimenter error, leaving 489 trials to analyze. (The means, standard deviations, and ranges of exploration times for each of the three conditions were reported in section 2.) A 3 condition (Visual Only, Haptic Only, and Visual and Haptic) by 8 stimuli ANOVA was conducted with the time spent exploring as the dependent measure. There was no difference in exploration time among the three priming conditions, (F2, 487 = 0.309, ns) or among the eight different stimuli (F7, 482 = 0.409, ns), and no interaction. Because the time spent exploring the objects was similar in all three conditions and across the eight different stimuli, differences found in object matching can be attributed to the different experiences afforded by the priming conditions. 3.2 Object matching (i) Will priming affect performance in a subsequent haptic-to-visual task resulting in a pattern of systematic matching behavior? Figure 3 shows the means with which children chose shape (white bars), texture (striped bars), and color (dotted bars) matches for the exemplars in each of the three conditions. The dashed line shows the chance level of performance (2.67) given three choices. Onesample t‑tests confirm the impression from figure 3 that shape choices in each condition (Visual Only: t20 = 6.46, p
