Novel Object Naming in Experiment 2

國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

69

familiarity will make the strengths become stronger so that the possibilities of other erroneous retrieval will decrease accordingly.

In this experiment, some facts were revealed. First, shape similarity indeed will fortify the strengths of the word to intervene children‟s word retrieval even though this kind of effect was lower for the familiar objects in Mandarin Chinese children.

Second, priming is a necessary factor to make children to give „target‟ error naming especially in the unfamiliar condition because the strengths for the unfamiliar objects should be weaker in children‟s mental lexicon. Third, word retrieval errors should not be totally affected by previously accessed words. Rather, concepts and words in our long term memory should play an important role. This is why there were many extraneous errors and why „control‟ error naming data almost completely came from children‟s labeling qiu2 „ball‟ to substitute for the things with round shape. From the data in our experiment, three types of error still could be explained by Dell‟s (1986) spreading activation model. Our finding is generally consistent with Gershkoff-Stowe et al‟s (2006) study.

4.4 Novel Object Naming in Experiment 2

Children‟s novel objects‟ naming results are shown in the table 7. The boldfaced numerals were the frequency of responses that happened due to the effects of target prime pictures in the corresponding condition where children participated (e.g.,

‧

labeling novel objects as high-similarity target prime in the high similarity condition).

Table 7. Children‟s Novel Objects‟ Naming Results Prime target

omission extraneous high low unrelated filler 2-year-old different ages in different conditions. In the high similarity condition, the responses that 2-year-old children gave most frequently were „extraneous‟ one (55.1%) while for the 4-year-old children, the responses were high-similarity target prime (40.6%). The responses of high-similarity target prime for the younger children were 26.1% of the time. And in this condition, both low- and unrelated-similarity target prime did not appear at all in the two different age groups. In the „omission‟ type, it happened more

frequently for the 4-year-old children (32.8%) than 2-year-old ones (17.4%).

„Extraneous‟ naming occurred 20.3% of the time for the 4-year-old children. And both

age groups seldom gave „filler‟ responses (1.4% for the 2-year-old children and 6.3%

for the 4-year-old children).

In the low condition, 2-year-old children were susceptible to the effects of the

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

71

target prime easier than 4-year-old children (age 2: 10.7%; age 4: 4.7%). And no matter for which age, they both produced the target names of the high-similarity condition (age 2: 10.7%; age 4: 14.1%). No children produced unrelated target primes for both age groups in the low-similarity condition. And the frequency of answering of „filler‟ was low both in 2-year-old (1.8%) and 4-year-old children (0%). For

children aged 2 in the low-similarity condition, the frequency of „omission‟ and

„extraneous‟ was 26.8% and 50% respectively, while for children aged 4, the

frequency was 43.8% and 37.5%.

In the unrelated condition, the possibilities of replying the unrelated target primes were very low both for children in two age groups (age 2: 1.8%; age 4: 0%).

Sometimes, they may give the target prime responses of high- or low-similarity conditions. For children aged 2, the frequency of answering high similarity targets was 8.9% while 0% for the low similarity targets. For children aged 4, the frequency of answering high similarity targets was 7.7% while 3.8% for the low similarity targets. No „filler‟ answers were produced when children labeled the novel objects in both age groups. Two-year-old children gave no responses 30.4% of the time while 4-year-old children gave no responses 25.6% of the time. Finally, the frequency of extraneous naming was close in both age groups (age 2: 58.6% vs. age 4: 62.8%).

As a whole, we could observe that target priming pictures would affect children

‧

on novel objects‟ labeling because there are perceptual similarity between them. And different levels of shape similarity have different levels of activation strengths on children‟s word retrieval no matter for children aged 2 or aged 4. Thus, we could find that with the decrease of shape similarity, the frequency of target prime responses was also getting lower. For example, for children aged 2, the frequency of target prime responses was from 26.1% to 10.7%, and finally to 1.8%. And for children aged 4, the situation was similar. The frequency was from 40.6% to 4.7%, and finally to 0%.

From this data, we could corroborate the effect of shape on children‟s word retrievals.

The priming from high-similarity concepts will strengthen the activation strength more than priming from low-similarity or even unrelated concepts. Besides, we used (2) age × (3) similarity AVOVA test to check the main effects on the frequency of children‟s „target‟ naming. And the results are shown in the table 8.

Table 8. The Effects of Age and Similarity on Children‟s Target Naming

Source DF Type I SS Mean Square F Value Pr > F age 1 0.01 0.01 0.01 0.9413 similarity 2 64.64 32.32 27.09 <.0001

From this table, we could know that only similarity was significant to target frequency (SS=64.64, MS=32.32, F(2, 43)=27.09, p<0.0001). Age did not approach significance (SS=0.01, MS=0.01, F(1, 43)=0.01, p=0.9413>0.05). And there was no interaction effect. Thus, we could corroborate that different levels of similarity should affect the frequency of children‟s „target‟ naming. High similarity indeed makes children

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

73

produced primed targets much easier than low similarity and unrelated similarity conditions. However, age seems not to make significant differences. That is to say, Mandarin children aged 2 and 4 both had the same performances on „target‟ naming.

This is not consistent with Gershkoff-Stowe et al‟s (2006) study. They had main effects of similarity and age at the same time. Two-year-old children were susceptible to the primed targets easier than 4-year-old children.

Although children in both age groups seem to be affected by primed targets, we still could observe some different influences between these two age groups. In the high similarity condition, the priming effect in older children appears to be stronger than younger children since they had 40.6% reply of targets while younger children only had 26.1%. However, in the low similarity condition, younger children seems to be affected by low-similarity targets (10.7%) easier than older children (4.7%).

Similarly, in the unrelated condition, the frequency of replying the name of unrelated targets in younger children was a little higher (1.8%) than the frequency in older children (0%). Older children appear to be affected easier only by high perceptually similar pictures than younger children while for those low or no perceptually similar pictures, they do not be affected that much. This phenomenon may be able to be explained by spreading activation mechanism as well (Dell, 1986). For children who speak Chinese, their mental lexicon‟s networks should develop from simple to

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

74

complicate. Since older children should have more complicated and complete networks, the concepts which are activated from one related node are likely to have more powerful strengths than the node in a more simple and incomplete networks in younger children‟s mental lexicon. Because of the immaturity of younger children‟s mental lexicon, the link between node and node should be weaker so that although the priming effect still exists in the high similarity condition, it cannot be that strong as for older children. Similarly, in the situation of low and unrelated similarity condition, since the linkage of nodes in older children‟s lexicon should be more mature, older children would not be affected by those conceptually farther primed targets. On the contrary, as the immaturity of younger children‟s mental lexicon, they would be affected by shape similarity or priming effects easier than the older ones. Thus, they had higher probabilities to reply the primed targets in the low and unrelated similarity conditions. Accordingly, as a whole, children in both age groups had no significant differences towards frequency of answering primed targets because both of them would be affected by shape similarity. However, with different levels of shape similarity, children still have different responses to them. This finding seems not to agree with our hypothesis where we assumed that 2-year-old children should be vulnerable to primed targets easier than 4-year-old children because there appears to be no differences between these two age groups from the ANOVA test. In fact, this

‧

result should be attributed to the fact that high similarity would fortify the strength of a node easier in a more mature mental lexicon than a more immature one, which results in the fact that both age groups have similar reactions towards the primed targets while it is not the same case in English. This finding may suggest that the way of words‟ processing or linkage between nodes and nodes in mental lexicon are different between Mandarin Chinese and English children. In addition, notably, our result does not mean that 4-year-old children are more susceptible to previously retrieved words because we did not observe the phenomenon that they are affected by low or no perceptually similar primed targets in the same way. Actually, from observing the responses of older children to the low similarity or unrelated primed targets, we still want to say that older children should be less susceptible to previously retrieved words.

To understand whether children will produce more extraneous naming when the degree of similarity decreases, a (2) age × (3) similarity ANOVA test was used to examine the effects of these two factors on children‟s extraneous naming. And the ANOVA results are displayed in the table 9.

Table 9. The Effects of Age and Similarity on the Frequency of Extraneous Naming

Source DF Type I SS Mean Square F Value Pr > F age 1 13.15 13.15 2.33 0.1342 similarity 2 37.03 18.51 3.28 0.0472

From this table, we could know that only similarity was a main effect on children‟s

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

76

extraneous naming (SS=37.03, MS=18.51, F(2, 43)=3.28, p=0.0472<0.05), but age did not reach significance again (SS=13.15, MS=13.15, F(1, 43)=2.33, p=0.1342>0.05).

And there was no interaction effect. Generally speaking, different ages did not make the frequency of extraneous naming differ significantly. However, in Gershkoff-Stowe et al. (2006), they still found main effects of age and similarity. In Mandarin Chinese data, children gave extraneous answers most frequently in the unrelated condition for both two age groups as we have predicted (age 2: 58.9%; age 4: 62.8%). And with the decrease of similarity, the frequency of extraneous naming was getting higher especially for older children (high: 20.3%; low: 37.5%; unrelated: 62.8%). The reason should be the same as we have mentioned before. That is, since older children have a more mature mental lexicon, the linkage between nodes and nodes should develop better than younger children‟s. Thus, they presented a more beautiful pattern for extraneous naming (i.e., less powerful priming effects result in more extraneous naming). On the contrary, due to a more scattered and loose linkage in younger children‟s mental lexicon, their pattern of extraneous naming (high: 55.1%; low: 50%;

unrelated: 58.9%) was not that beautiful as older ones‟. Despite of this, the extraneous naming still happened most frequently in the unrelated condition. Due to the possible different way of processing and word retrievals between Mandarin and English, Mandarin children in two age groups did not have significantly different extraneous

‧ 國

立 政 治 大 學

‧

N a tio na

l C h engchi U ni ve rs it y

77

naming. Older children in English seem to be less susceptible to the effect of previously retrieved words because of their mature mental lexicon. This fact causes them to activate more extraneous naming. However, more mature mental lexicon in Mandarin Chinese children seems to make them have more elaborate lexical networks so that this contrarily make the effects of conceptually related prime become even stronger. Accordingly, there was no age effect on children‟s extraneous naming in Mandarin Chinese children.