【作者】 胡林成；

【导师】 熊哲宏；

【作者基本信息】 华东师范大学 ， 基础心理学， 2009， 博士

【摘要】 数量模块的理论与实验研究二十世纪80年代，福多在《心理模块性》中界定了认知心理模块的构成标准。二十世纪80年代末，进化心理学诞生，进化心理学家提出了“达尔文模块”的思想。“模块心理学”在上述两种模块思想的基础上得以初步建构，这使得数量模块的研究成为可能和必然。本研究旨在从理论和实证两个路径论证数量模块存在的可能性。数量模块的研究不仅能够为已有的各种数量加工方面的实证研究结果提供理论解释，也能够为理解和进一步研究人类的数量加工能力提供合理的视角。本研究分为两个部分，第一部分是理论研究，其研究目的是论证数量模块存在的可能性。该部分以模块的三个构成标准为主线，从已有的数量加工的研究文献入手，对数量加工的相关研究结果进行分析和综述，用数量模块假设解释已有的相关发现。模块的第一个标准是天赋性。对人类的数量加工能力的研究表明，成人、婴儿和非人灵长类动物在数的表征和简单的数推理方面具有共同性，人类的数量能力可以在进化历史中找到源头，这有力地证明了数量模块的天赋性。模块的第二个标准是领域特殊性。越来越多的证据表明，数量能力有其不依赖语言的演化基础。算术障碍儿童的研究成果以及脑损伤的研究结果同样表明，数量能力独立于其它的认知能力。脑成像研究也表明，人类的数加工具有特定的脑区，这进一步为数量加工能力的领域特殊性提供了生理基础。模块的第三个标准是信息的封闭性。目前研究者热衷的SNARC效应和数字Stroop效应都可以用数量模块的信息封闭性假设来合理解释。另外，对数量加工的无意识研究也为数量加工的信息封闭性提供了有力的证据。以刺激材料产生的数量感类型为标准，数量模块相应地划分为数字子模块、模拟量子模块和离散量子模块。在理论研究部分发现，虽然已有的研究结果支持数量模块假设，但是，这些研究基本局限于数字模块方面，主要是对阿拉伯数字、数词以及其它与数有关的语言材料上，其主要发现是SNARC效应、Stroop效应、距离效应以及启动效应。对模拟量模块和离散量模块的研究则非常少。因此本研究的实验研究的重点是对模拟量和离散量的加工特点进行研究。本研究的关键假设是，数字模块在加工数量信息时所表现出的特点(如SNARC效应、Stroop效应、距离效应以及启动效应等)可能是进化序列中更为基础的离散量模块和模拟量模块所具有的特点。从数量模块的发展角度来看，数字模块加工的上述特点在儿童接受大量的数学训练之前可能并不存在，而模拟量模块和离散量模块有可能在儿童接受大量的数学训练之前就表现出上述特点。本研究的第二部分是实证研究。实证研究分为五个部分，研究一用面积和亮度作为比较刺激来考察存在于数字比较中的SNARC效应是否存在于模拟量的比较任务中(如面积比较和亮度比较任务)。研究二探讨在数字的物理大小(即模拟量)的比较任务中是否存在启动效应，数字物理大小的无意识比较任务中是否存在类SNARC效应。研究三探讨数字个数和数字语义之间是否存在类Stroop效应，在个数(即离散量)比较任务中是否存在类SNARC效应。研究四以尚未接受数学技能训练的5岁左右幼儿为被试，采用数字大小比较和面积(即模拟量)比较两类任务，以SANRC效应以及距离效应为指标来测试5岁左右的幼儿是否能够对数量进行空间表征。研究五用数字个数(离散量)作为比较任务，来考察5岁幼儿在个数比较任务中能否产生个数数量的空间编码。实验研究的结果表明，数字加工中存在的SNARC效应、距离效应、启动效应和Stroop效应均存在于模拟量比较和离散量比较任务中。而且，尚未接受数学技能训练的5岁被试在离散量和模拟量比较任务中发现了SNARC效应，而在数字比较任务中则没有发现SNARC效应。在总讨论部分，用数量模块的三个特性对实验研究的结果进行了详尽合理的解释，并且在已有的解释SNARC效应模型的基础上提出了数量模块内部加工机制的新模型。更多还原

【Abstract】 A theoretical and experimental Study on Magnitude ModuleIn the 1980’s,J.Fodor put forward the standards of module.In the late of 1980’s,evolutionary psychology appeared and Darwinian module was proposed byevolutionary psychologists.In 2004,Module Psychology was put forward bycombining Fodorian module and Darwinian module.The idea of module psychologymade the study of magnitude module to be possible and necessary.The current studyaims at demonstrating that the hypothesis of magnitude module is possible by the wayof theoretic reasoning and experimental study.The study of magnitude module couldoffer theoretic interpretation for existing experimental findings,and throw light on theunderstanding of magnitude competence.The current study is divided into two parts.The first is to reason the possibilityof magnitude module resorting to the existed findings in the aspect of magnitudeprocessing.In detail,the first part of current study intend to reinterpret the existedfindings by analysing and summarizing the existed findings according to the threeingredients of module.The first essential ingredient of module is inherence.The study on the magnitudecompetence suggested that human infants,human adults and non-human primateshare many common characteristics,and human being’s magnitude competence couldtraced back to the evolutionary origins,which argued that inherence is one of theessential ingredients of magnitude module.The second essential ingredient of moduleis domain-specific.More and more findings suggested that manitude competencepossesses of its evolutionary foundation that is independent of linguistic ability.Manyneuropsychological studies of adults with neurological damage strongly indicatc thatnumber knowledge is dissociable from semantic memory (Cappelletti,Butterworth,&Kopelman,2001),and that the semantic memory systems for numerical and non-numerical information are localised in different areas of the brain (Thioux,Seron,& Pesenti,1999).Functional neuroimaging reveals that the parietal lobes,especiallythe intraparietal sulci,are active in numerical processing and arithmetic (Dehaene,Piazza,Pinel,& Cohen,2003),and studies of brainlesioned patients (Cipolotti & vanHarskamp,2001) have identified the left lntraParietal Sulcus (IPS) and the angulargyrus as critical to normal arithmetical performance.These findings indicate thatmagnitude competence is domain-specific.The third essential ingredient of module isencapsulation.The results of number recognise researches such as SNARC effect,Stroop effect could be interpreted by the encapsulation of magnitude module.Additionally,study results about unconscious number processing provide the evidencefor the encapsulation of magnitude module.Magnitude module is divided into number module,analog magnitude mdule anddiscrete magnitude mdule according to the categories of number sense that the stimulicaused.In theoretic part,it is revealed that the existing findings support thehypothesis of magnitude module,however,most of existing studies focused on thenumber module,and the main findings are SNARCeffect,Stroop effect,distanceeffect and priming effect of number.Analog magnitude mdule and discrete magnitudemdule are not explored extensively.Hence,the current study is focused on analogmagnitude mdule and discrete magnitude mdule.The key hypothesis of current study is that the characteristics that numbrermodule manifested when processing number are originated from analog magnitudemdule and discrete magnitude mdule.From developmental perspective of magnitudemodule,the characteristics that numbrer module manifested when processing numberare not existed before children accepte extensive arithmetic training,but themagnitude process of analog magnitude mdule and discrete magnitude mdule mightpossess the characteristics above.The second part of current study is experimental study,which consists of fivestudies.Study 1 aimed at exploring whether SNARC effect exists in area comparisontask and luminance comparison task.In a binary response setting,it has been frequently observed that small numbers are responded to faster with the left hand andlarge numbers with the right hand (i.e.the SNARC effect) which reflects the spatialleft-right orientation of the mental number line (Dehaene,Bossini,& Giraux,1993).So it was worth considering whether spatial associations are exclusively numerical orwhether they could occur with continuous properties of non-numerical stimuli (e.g.area,luminance,hue of a figure;length,width and weight of an object;pitch andloudness of sound).In two experiments,the area and luminance of circle were used ascomparison stimuli to test the hypothesis that not only number comparison task butalso the comparison task of continuous properties of non-symbolic stimulus couldproduce SNARC effect.Both experiment 1 and experiment 2 used numbercomparison task as the index of producing SNARC effect,and the results proved thattwo modes of stimuli presentation were valid.In experiment 1,the area of circle and the semantic of number were used ascomparison stimuli to test our hypothesis.The results indicated that left handresponded faster and more accurately to relative small magnitude of stimuli (area ofcircle or semantic of number) and right hand is faster and more accurate to relativebig magnitude (area of circle or semantic of number).SNARC effect of numbercomparison task in experiment 1 conforms to the study of Dehaene etc.(1993).Inarea comparison task of our experiment 1,bigger circles were answered about 47 msfaster to the right than to the left,and smaller circles were answered about 12 msfaster to the left than to the right.In experiment 2,to explore whether the luminance of stimuli could produceSNARC effect that was found in number comparison task,the luminance of circle wasused as comparison stimuli to test our hypothesis.For luminance comparison weformed five circles with different luminance (from 0 to 255 in RGB -coded shades ofgray) that varied only in brightness with constant hue and saturation.The resultsproved that both number comparison task and luminance comparison task couldproduce SNARC effect.Left hand responded faster to higher luminance circle andright hand responded faster to lower luminance circle.Conclusively,the results of study 1 indicated that SNARC effect could be obtained in area comparison task and luminance comparison task.Furthermore,theresults suggested that (1) distance effect that was observed in number comparison taskalso existed in area comparison and luminance comparison,Stroop effect that wasobserved in number comparison task also existed in area comparison,(2) the responsespeed to area comparison and to luminance comparison was faster than to numericalcomparison.It is found that number comparison task and physical stimuli comparison taskshare the same process mechanism(Gallistel & Gelman,2000).Unconscionsly,Dehaene,Naccache et al.,(1998)and Naccache & Dehaene(2001) revealed that theunconscious semantic priming exists in number process.Hence,in study 2 we inferthat unconscious process might exist in the process of number physical properties.The current study aims at whether priming effect and SNARC-like effect exist in thephysical properties comparison task.Using masked priming procedure (Naccache &Dehaene,2001),in both number comparison task and number physical sizecomparison task,experiment 1 found SNARC effect,priming effect and Stroop effectin 33ms priming condition.Experiment 2 found SNARC effect,priming effect andStroop effect exist in both number comparison task and number cover areacomparison task in 33ms priming condition.Numerical Stroop paradigm revealed that the physical size of number and thenumber of numerals (small magnitude) could automatically activate semanticrepresentation,hence produce interference and facilitation between physical size,number of numerals and number magnitude.Based on this automatic process,weinfer in study 3 that in the task of number(i.e.,quantity of units or individuals)approximate estimation might activate automatically the spacial representation ofmagnitude of number.The current study designed two experiments to test thehypothesis.By using Stroop-like effect as index,the first one explored whether thenumbers of numeral and number magnitude could produce interference andfacilitation;the other experiment explored whether the number(i.e.,quantity ofasterisks) could represent spacially by using SNARC-like effect as index.The resultsrevealed that there exists Stroop-like effect between number of numerals and number magnitude,and there exists SNARC-like effect in the number(i.e.,quantity ofasterisks) comparison task.We also found an order effect,that is to say,when themagnitude of numerals number of standard was smaller than that of comparisonsimuli,Rts were faster,otherwise Rts were slower.SNARC effect can be influenced by children’ growing numerical skills and taskdemands.By using number comprison task and area comparison task that are releventto magnitude,in study 4 we test whether 5 year-olds could automatically activates aspatial representation when representing magnitude of Arabic number and magnitudeof circle area.Rts data of number comprison task analysis indicated that there is noSNARC effect,however,distance effect is found.Error rate analysis indicated thatthere is no SNARC effect or distance effect.Rts data of circle area comprison taskanalysis indicated that there are SNARC effect and distance effect.Relatively largecircles are responded to faster with a right response than with a left response andrelatively small circles are responded to faster with a left response than with a rightresponse.Error rate analysis indicated that there exists SNARC effect but no distanceeffect.By using the task of number comprison of numerals,we test whether 5 year-oldscould automatically activate a spatial representation when representing number ofnumerals in study 5.Rts data analysis indicated that there is SNARC-like effect,however the orientation of the SNARC effect was reversed.Relatively more numeralsare responded to faster with a left response than with a right response and relativelyless numerals are responded to faster with a right response than with a left responsc.Stroop effect was not found,which means that number comprison of numerals did notactivate automatically magnitude of numerals.Error rate analysis indicated that thereis no SNARC effect or Stroop effect.The results of five experimental studies suggested that SNARC effect,Stroopeffect,distance effect and priming effect of number when processing number alsoexist in the process of analog magnitude comparison task and discrete magnitudccomparison task.Furthermore,SNARC effect is found in the task of analogmagnitude comparison and discrete magnitude comparison when subjects arc five-year-olds,however in the task of number comparison,five-year-olds do not showSNARC effect.The experimental results support the key hypothesis of current study.In general disscussion,experimental results are interpreted intensively andrationally according to the three ingredients of magnitude module.Further more,anew model about the mechanism of magnitude module is proposed on the basis ofexisted model that explains SNARC effect.更多还原