A STUDY OF CONCEPT FORMATION IN CHILDRENl
CHIEN-WEN
Su
1. INTRODUCTIONThe term “concept" has been defined by McDona娟的“acIassification or systematic
organization of stimuli."(6) Therefore
,
a concept is not direct sensory data but derived from the elaboration and combination of sensory data. The common elements in discrete objects or situations serve to unite object怒。r situations into a concept. The attainment of certain concepts actual1y 1控制的 that one has 為cquired some tools to understand his environment and some degree of mastery over it.How does a concept develop? As des位ibed by Hul1, "Concept formation is a process of generalizing abstraction which involves the discrimination of a particular element common to variety of stimuli.叫 3) 1n this light
,
there are two important proc您認es involved in learning a concept.(主) A discrimination is required. That 誨， one must be able to distinguish the characteristics of a given object or situation from the others and to know the similaries and differences of them. For example, if the child knows what a circIe 誨， he must be able to distingui也 circIes from squares, triangles or other forms. If he fails in doing so, we will doubt that he has the concept of roundness at aIl.
(主) A generalization is required. GeneraliZation is avery complic乳ted thinking process which involves recognition of the present stimu1間.pattern and refl.ective thinking in order to relate the present stim泌的-pattern with the old experiences. He must be able to utilize the knowledge formerly acquired to interpret new ex-P釘iences. This indicates that one must not only have the abstracting ability to grasp the common elements from discrete experiences but also can use them as basis to categorize or to group objects.
How do children form basic concepts according to the above mentioned prin-ciples? What are the important characteristics of concept formation in children of different ages? Th部e questions are my maior interest of concern. Specifical1y speaking, the main purpose of this study is to use forms and colors as stimulus-patterns to find out the fol1owing characteristics:
(工) The concept of forms and colors of chi1dren of di品rent ages. (芷) The discriminative ability of children of different ag!的.
(草[) The generaIizing abi1ity of chi1dren of different ages.
(IV) The important characteristics of grouping behavior of children of different
ages.
1 This study was subsidized by the National Council on Science Development.
B西
11. PROCEDURE
(工) The Subject: 55 chi1dren are selectd for this study. Among them, sixteen gir1s and twenty boys are kindergarten children‘ Six girIs and thirteen boys are 我rst and s忱。nd graders of a pub1ic primary school. They are selected at random from the enroIIment of the school year of 1965可1966. Their age.range is 23 months to 96 month. The age.mean is 63 months
,
and st乳ndard deviation (SD) 21 month就 The subjects are grouped according to their chronological ages (CA). Children of two years are divided into two groups for the convenience of analysis. The IQ雕range of the subject is from 59.7 to 161.1 with a mean of 98.1, and its SD of 22.54. The average duration of their attendance at kindergarten is around one year. The first and second graders, inc1uding the years of their attendance at kindergarten, have been in school for three to four years. The personaI and family backgrounds of alI subjects acquired from interview with the parents are presented in Tables 1 and 11.Table 1. The Subjects Backgrounds
55 17
Table
rr.
The Subjects' Family Backgrounds(立) Materials: Five tests are used for this study. They are form discrimina-tion t臼t; color discrimination test; Goldstein.Scheerer color.form sorting test(B); Gelb.Goldstein color sorting test∞ and E. Hanfmann and K. Kasanin's concept formation test. (3)
{班) Procedure: The tests are given individualIy one at a time. The timing of the experiment is not arbitrariIy scheduled due to the short attention span of young children. The criteria for passing aII tests would be three correct responses made by the subject successively. If he faiIs to make cortect responses in six ∞n tinuous trials, the subject is considered having faHed once. If three such 如ilures are made by a 叫，bject succe部ively， he is out of the test. But
,
he wiIl have the chance to 投y on the next 胎st. Running records of subjects' respon甜s toward the tests are taken carefully and in great detaiI.AlI the kindergarten childr如何ke Merril-Palmer Scale, and the elementary school children take Binet-Simon Intelligence Scale. All the tests are given indivi. dual1y. The test scores of MerriI♂almer Scale are translated into IQ scores for the convenience of analysis.
As the regular experiments and intel1igence tests have been completed, the ex-perimenter goes to visit subjects' fami1ies to interview their parents in order to get more information about thè subjects and their family backgrounds.
ill. RESULTS
The focus of attention in this study is directed toward a comparison between various age groups in terms of correct responses to 由.e above mentioned five tests. The result wi11 be discussed in the foI1owing:
(1) Test One
The number of cases and their percentages of Ss in each age group who have made correct responses on 位le form discrimination test, on namingthe form and on giving correct verbal explanation for their choic器， are shown in Table ][.
The data reveals that the subjects' abi1ity to discriminate different forms appears
r祕ler e紅妙. 50 per cent of subjects (Ss) at CA 2-2.5 can discriminate circ1es and triangles; 75 per cent of them can discriminate squares. None of 泣lem under CA 2 has made any correct response, and 甘10se above CA 2.5 find no difficulty in making correct r的ponses to circ1es
,
squares,
and triangles.T嗨Ie ][ also shows that Ss younger than CA 3 are unable to name the various forms correctly. Ss of CA 3 and 4 know circ1e better than square and triangle. Reaching age 5, Ss can name the circIe, square, and triangle precisely.
Ss' ability to give correct verbal explanation for their choices begins to appear at CA 3, and 能is ability increases with age. Ss of CA 7 and above have made 100
per cent correct verbaI explanation for their choices.
Most Ss at CA 2-6 take circ1e as their first preference rather than sq關，re and triangle. Ss at CA 7 take triangle as their choice. The order of their preference is triangle (44.4是，%)， square (33.3~局)， and circIe (22.22%).
(11) Test Two
The number of ca揖s and their percentages of Ss in each age group who have made correct responses on 位le color discrimination test and succeeded in naming the color and giving correct verbaI explanation for their choices, are shown in Table IV.
Table IV shows that 50 per cent of Ss under CA 2 are able to make correct respon肥s to red, yellow, and blue colors. 50 per cent of Ss at CA 2-2.5 can d做到 minate red color. Only 25 per cent of them can discriminate yellow color but none of them can discriminate blue and green colors. Ss at CA 2.5--3 can discriminate red and yellow colors and only one third of them can discriminate blue and green colors. Up to CA 3 and above, Ss find no difficu1ty in discriminating red, yellow, blue
,
and green colors.CA忙軍重
lkJ!
盟軍罪証令世軍括全軍
Under 2 2ω 么 5 2.5 伽 3 3 是 56 7XJ
佳話:在
lEEK
直且必注
1
而
.reen
主
當Table 1V also shows that the earHest age for giving correct names of various colors is at CA 2-2.5. The age of mastery over it 誌 at CA 6. The Ss' earliest age for giving correct verbal explanation for their choices is at CA 3. The age of m部tery over it is 成 CA 7.
Table 1V reveals that there is a clear tendency in you.nger Ss toward taking red color as their first preference. 50 per cent of Ss at CA 2. 75 per cent at CA 2-2.5, 66.67 per cent at CA 2.5-3 are inc1ined to red color. After CA 3, along with growth. the preference for red color discreases continuaIly. Up to CA 7, there is only 20 per cent of Ss who prefer red color. Percentages of first preference scatter around for the rest of the colors. No specific preference is shown.
( 111) Test Three
Table V shows the resuIts of Test Three made by Ss. In this test, each S has taken three trials. The frequencies of correct r的ponses made by Ss either according to forms or to colors are ∞unted and their percentages are calculated. Table V reveals that Ss youngerthan CA 2 are unable to make correct r部ponses. These results indicate that before CA 2, Ss can hardly abstract a common element out of the various discrete objects.
Table V. Number and percentage of correct responses made by Ss at each of 8 age groups on color.form sorting te仗
Failure % 1∞ 75 11.11 1 16.66 55.56 58.33 37.04 50 62.5 93.33 2 5 40258 .ip 唔， AtA 穹忌的 4 。 9762 8.33 33.33 何制品間的 札榕的甜甜& 。 dnv 句 4 旬，卻 AM9 自 嚕，主唔， •• 啥霉，為 Under 2 2-2.5 2.5-3 3 4 wbhhv 旬，
In comparing the percentages of correct respons部 in grouping 怯e stimulus blocks according to forms and to colors made by Ss of each age grouP. we find a clear tendency that Ss at CA 2 and 3 當e apt to group them according to forn的
After CA 3, the percentage of groupings according to forms decreases while the percentage to colors increas能 At CA 4, SS.seem more likely to group the stimulus blocks according to colors. From then 0泣， the tendency of grouping according to colors decreases steadi1y. The dominant reaction pattern is shifting back to forms again at CA 5. 62.5 per cent and 93.33 per cent of the groupings made by Ss at CA 6 部被 7 are according to forms. (Figure 1) The four failures appeared at CA 6
00 % 100 會8 思8 70 60 50 40 30 20 10 。
戶 7
2 3 4 5 § 偏!幽耐輛輛_.峭 -Forrrt 一-一-一咱可 Color ? (CA)Fig. 1. The percentages of correct groupin醉 according to forms and to colors by ss in 開ch age group.
may be due either to the misunderstanding about the statement made by the experi機 menter or the nervousness of the subjects during the experiment.
As far as the verbal explanation for their groupings is concerned, Ss at CA 2 are unable to give any ∞rrect responses. This ability is shown at CA 3 only for forms. The percentage is 21.是3. The percentage發 of correct responses of Ss at CA 是 and 5 are 52.9是 and 58.33 for colo詣， and 60 and 66.67 for forms. From CA 6 0紋， Ss find no di益culty in giving verbal explanation for their groupings both for fon閑 and colors.
(IV) Test Four
The results of 寸的t Four are analyzed by counting the total number of groupings made by Ss of each age group. Their grouping behavior can be cla部ified into two categories: the concrete and the abstract approach. By using the concrete 叩proach， the subject tends to do the matching for identical color skeins instead of cla部ifying the skeins into four color 紋。ups (the color of red, yeIlow, bl啦， and green). The
matching procedure 1豆瓣結 certain range which depends upon the degree of concrete-n部s. The 缸· grouping is analyzed according to the extreme degree of concreten部3… matching for identicaI color skeins-and the lesser degree of concreteness. The latter means that the 缸， grouping involves more than identical color skeins but the number of them is fewer than that of the grouping which is considered as formulating conceptuaI thinking.
The abstract appro必ch means that the S consciously treats the skein 翁s a 帥" spr.晶entativeof a c前tain basic color hue. Although he knows the specified attribut帥
of an individual skein such as intensity, brightness and purity, he is able to ignore
怯的e differences and group them together according to a color concept. When asked the reason why he considers those skeins as belonging together, he is able to name the color category to back up his choice.
Table VI gives us a general idea how Ss respond to Test Four. It reveals that the average grouping made by Ss of each age group is more than four groups. The maximum reaches up to 10 at CA 壘. The average grouping decreases slightly along with age but these is another increase beginning at CA 6. Ss under CA 2.5 are una1;lle to make any correct grouPing. Their average grouping is from 1 to 1.75. This indicates that they just make one mixed groUP by putting all di仔erent ∞lor
skeins with various bright泣的描s into one group. A great amount of apprehension is shown.
Table VI. Average number and range of groupings, number and percentage of Ss,
at each of 8 age groups according to concrete 泛nd
abstract approach on color sorting test
Items: Number of
I
Concrete Approach Abstract Approach|…向前greE 叩叭的Of~1
Concretene串串 Concretene串串 RedY山
Gre
Blueed I Yellow ' Green
|缸恥tshrm
MMS1
CA \ IAVeragelRangel N
I
;;6 !GrOUPingsj 一N …一F4-Grouping器 N I N I N NUnder 21 1 1 。 O 。 2叫2.5 i 1.75 1但是 Oi 。 1 25 4 2.5-3 6 1-13 l 33.3 13 1 33.3 4 3 8.13 1-19 6 75 39 4 是。 17 1 1 4 10.44 7你 14 a 100 52 a 100 36 1 5 5.88 4明§ 4 4也是4 9 5 55.56 2是 2 2 2 2 6 8 4-12 是 50 22 3 37.5 13 2 2 3 3 7 10.1 4鵬主2 7 70 63 2 20 14 3 3 3 3
Generally speaking
,
the dominant reaction pattern made by Ss tow泌'd Test FOUI is concrete 部)roach. Matching for both extreme degree 我nd le部er degree of con“ creteness has appeared in each age group above CA 2.5. The percentages of Ss who have matched for identical color are 33.33 at CA 2.恥3， 75 at CA 3, 100 at CA 莓， 44.44 at CA 5, 50 at CA 6 and 75 at CA 7. The percentages of Ss of each 為gegroup who have 混atched for lesser degree of concreteness are 33.33, 50, 1∞， 55.嗨，
37.5 and 20 respectively. In comparing the percentages of Ss who have matched for identical color with th的e who have matched for le部er degree of concretene錯， we find that Ss in each age group are inc1ined to match for identical color. This means that they take the extreme degree of concrete approach 鍋 the basis of their groupmg.
如
There are only a few Ss who have shown the tendency of taking the abstrad approach in 位leir groupings. The S in this study who first showed this attitude is a girI of three years and ten months of age. She is 紅ble to group aII skeins with different shades and brightnesses in red and yelIow according to the color concept but has failed to do the same in blue and green colors. She can even teIl the ex-perimenter the correct verbal explanation for her choices. Of the 55 齡， there are only 10 Ss who can produce groupings according to 設le abstract approach, one at CA 3, one 拭 CA 巷， two at CA 5，位lree at each of CA 6 and 7.
(V) Test Five
The concept formation test used for this study is a rather compIicated one which requires not only discriminative ability, ability of sorting blocks according to their common characteristics, but also abiIity of making gen位alizations according to old experiences. Careful observatiol1 is made, and errors made by Ss for 粉ch of the
Lag, Mur, Bik and Cev 開tegori的 are counted. The average frequency and range of errors made by Ss of each age group are presented in 寸的le VH.
XMK!正hyltht在
;lf
“明mu m--呵 ，i 唔， anu iiidqin 是吟，如本 唔， A-z-a whM HI制 唔，‘ 旬， AF IPFOS quI 品， A AVLAnr 4z •• gd' .... h a3& 啥 ••• , .. h 11.4 I 0-23 是 10-14 1 I 0幽9Table VH shows that Ss under CA 2 are not able to handle the task comprehen. sively. Their typical reaction is just grasping the bCocks in hand to play or refusing to act accordingly. All of them fail to carry out the performances.
Observing the average and range of errors made by Ss at CA from 2 through 6 for each of the Lag, Mur, Bik, and Cev categori嗨， we can understand that Ss' grouping proc制制s are completely based on trial and error behavior. No insight whatsoever is evidenced. AccidentaIIy, some Ss at CA 是 and 5 have found that there are certain marks on the bottom of the blocks. They start to look for the marks and use the訟部 the basis for their groupings. However, they can not under-stand there is a meaningful connection between the mark and the blocks.
Insight and generalizing abi1ity are first shown at CA 7. This is exhibited by a remarkable decrea峙。f errors made by Ss for each of the Lag, Mur, B此， and Cev categories. This índicates that Ss younger than CA 7 are unable to grasp the meaníng of the task of classification. They have no insight intQ the multiple po岱L bi1ities of grouping the blocks but try to match them eÍther according to colors or to forms. They can never actively apply definite categories to the materiaI but are rather passively led by vaguely perceived simiIarities of the blocks which may become related to one another as individuals. This tendency is justifie技 by Ss' verbaI explanations which indicate that they can only grasp partiaUy the characteristics of the blocks. Their answer海 are either large or taU, smal1 or fl鼠， but never combine
both together.
IV. CONCLUSION AND DISCUSSION
The results of this study provide us with some understanding about the nature of the proce紛 of form and color concept formation in chiIdren of different ages. Judging from the data shown in tabl的車， N, V, VI and
vn
, we know that Ss under CA 2 can hardly make any correct responses toward the five tests used in this study. It seems that the third ye我r of life is an important period at which chiIdren start to construct the outside world according to their common elements.The 郎· discriminative abiIities for forms and ∞lors appear in a paralIel fashion at two years of age. But, their 品i1ities to discriminate vario臨 forms or ∞lors do not appear at the same time. This study shows that a single form or color is named correctly whi1e other forms or colors are sti1I confused. Ss under three predominantly u總 form in preference to color as the basis for sorting the blocks. Between the age of three and five there is a gradual shift to color; but after five form becomes dominant again.
Both form 詠nd color matching antedate form and color naming which in turn appears prior to giving correct verbaI explanations for their choices. This fact in-dicates that Ss' grouping behavior is from percePtuaI to conceptuaI leveI. The b部is of grouping for Ss younger than CA 3 is solely dependent upon the sensory cohesiv心
如臨 of impression. This is only a primitive and lower level of discrimin叫íon. At CA 3 and above
,
the Ss can name the color or the form correctly. Therefore,
we may consider them as having got the meaningful und位standingof the form or color and having acquired the concept of it. In this Iight, we can conclude that Ss' 的ility to make generaIizations first appears 的 CA 3. From then on, there is a st給你in位ease in their abiIity to group objects together and to give abstract and conceptual explanations of the groupings.
Ac∞，rding to 由e data of Tables VI and 翅， most Ss in this study have taken the concrete attitude in groupÍng and sorting things. That means Ss tend to do the matching for the identical color instead of c1assifying the skeins into four color categories (red, yelIow, blue, and green). Their matching for the identicaI color
94
skeins or forms is due to their treating the stimulus-pattern as a thing and not as a representative of a certain category. The bases of their groupings are completely dependent upon the perceptible charact訂istics of the stimulus-pattern. This fact is 訓 expression of Ss' lack of a gen宙al reference and experiences. This tendency has also re色的ed it田lf in their reaction patterns of hesitancy, in calling skeins of various shades of one color by di任erent nam喲， and in n雄kingpseudo serie就 Pseudo series means that some Ss can produce a grouping which includes quite 統 large number of skeins belonging to 認 certain basic color concept. But, when the experi-menter asks the reason why those skeins belong together, the subject wi1l clai詛 that the skeins which he has formerly grouped together are not the same. 盯sually， he would re紅range them according to the identkal color or would give them different naI瓷器. We, therefore, understand that the subject has n。如ch categorical point of view in mind nor does he have the intention of forming a 阻ries according to a color concept.
Data shown in Table VII wilI give us a more clear心的 picture about the fact that 位le Ss younger than CA 7 are unable to assume the conceptual or the abstract attitude to the tas丸 These re純lts will support the findings of Test Four. Com-pliances with the grouping or sorting ta成sof Tests Four and Five definitely requir臼 three most importaJ泌的ilities. They 認re the categorical attitude, the insight into the multiple possibi1iti的 of choice, and the consideration of 位le whole. The cate-gorical attitude embraces more than the reaI stimulus in its scope. It∞nsis惱， in our setting, of viewing the stumuI帥-pattern not as individuals but as bearers of certain general characteristics and as representatives of a certain category. The majority of Ss in this study lack 也is attitude. Con甜quentIy， Ss are limited to the percePtible characteristics of tl唸 stimulus.pattern. This is the reason why in Test Four, they group the identical color skeins together and in Te斜 Five， they 豆nd no way to group the blocks togeth叮 except turning them up to read the marks accord. ing to which the groupings are 認錯成 The second aspect, closely related to the 位rst one, is the insight into 技le multiple possibilities of the choice. This abi1ity is especially required of Ss in order to perform Test Five correctly. A subject who lacks this insight may be able to organize the
REFERENCES
( 1) CARPENTER F.: Conceptualization as a function of differential reinforcement, Science
&/:都catåon， 1954， 錯:， 2棍-94.
(2) DAS腿也..L ]. F.: Abstract and concrete behavior; an experimental study with special tests,
Psychological Monogra余份， 1941.
(3) 旦 ARRIS T. 泣. and SCHW AHM W. E.: Selected readings 仰 the learning proc憫， New York,
Oxford Univ. Press, 1961.
(4) Hu技LOCK EUZABETH 怠.: Child devel，。如耐， Graduate School of Education, Univ. of Pennsylvan泣， 1奶是.
(5) KENDLER T. S.: Co終cept formation, Annual Review 01 Psychology, 1962， 訟.4是7刪72.
( 6) MCDONALD F.].: Educational Psycholo，鯽" Wadsworth Publishing 仇， 1帆. San Francisc心 1959.
( 7) OSEAS L. and UNDERW.∞<D B.J.: Studies of distributive practice: V. Learning and reten-tion of concepts. JOU1糊101 E:rþ3rimental Psychology, 1952， 她 143-48.
(8) VAUGHAN G. M.: Concept formation and the development of ethnic awareness, The Journal
olGe雕tic Psycholo，鯽" 1奴詣， 1(娥的-103.
(9) WHITMAR制 G.A. and Others. A study of developmental changes in conceptual and perceptual associative c1ustering. Journal 01 Genetic Psychology, 1紡串，魄的心位
(10) WOHLWILL ].].: The abstraction and conceptualization 01 form, color, and number. Journal
、帽、 搗
