AQA心理 感觉知觉 深度线索 错觉 两种理论
Perception 只是 AQA GCSE Psychology Paper 1 里的 Section B，25 分，题目却年年围绕同一批概念打转。感觉与知觉的分界、四类深度线索、两种对立的理论、四个影响因素，构成了这 25 分的基本盘。下面的内容按 AQA 官网 specification 第 3.1.2 节的顺序梳理，每一节都可以直接拆成答题句。
Perception sits in Section B of Paper 1, worth 25 of the paper's 100 marks, and the questions rarely wander far from the same handful of ideas. The distinction between sensation and perception, four kinds of depth cue, two rival theories and four factors that shift what we see make up almost every question set. The notes below follow the order of section 3.1.2 on the AQA website, so each heading lines up with the specification you are examined against.
感觉与知觉的分界
感觉是感受器接收外界物理能量的过程，眼睛收到光，耳朵收到声波，这一步不需要理解。知觉是大脑对这些信号的组织与解释，它把两维的视网膜图像变成一个有距离、有物体的世界。考试里区分两者最短的说法是：感觉负责检测，知觉负责解读。很多学生以为知觉只是感觉往后延了一步，其实两者在时间上几乎重合，区别在于处理层次不同。看到一个红色圆形物体，感觉层面只有颜色与轮廓，知觉层面才出现苹果这个概念。AQA 官网的 specification 把这一条放在 3.1.2 的第一行，因为后面所有内容都建立在它上面。
Sensation is the detection of physical energy by receptors in the eye and ear, and it requires no understanding; perception is the brain organising and interpreting those signals into a world of objects at a distance. The shortest exam-ready distinction is that sensation detects while perception decodes. A red round shape arrives as colour and contour, and only at the perceptual level does it become an apple.
单眼深度线索
单眼线索用一只眼睛就能获得，AQA 的 specification 列出四种：高度、相对大小、遮挡与线性透视。高度指物体在视野里越靠上就被判断为越远，靠近地平线的东西都落在画面上方。相对大小依赖我们对物体真实尺寸的经验，两个人看起来一样高时，画面里那个更小的通常站得更远。遮挡最直接，被挡住的东西在后，挡住别人的东西在前。线性透视来自平行线在远处汇聚，铁轨与公路都给出这种收敛的图形。常见的错误是把相对大小说成近大远小，线索描述的其实是同一物体在不同距离上的成像变化。
Monocular cues work with one eye closed, and the specification names four: height in plane, relative size, occlusion and linear perspective. Occlusion is the least ambiguous, because a covered object must lie behind whatever covers it; relative size relies on knowing roughly how large things really are. Parallel lines converging towards a vanishing point give the same depth information as railway tracks receding into the distance.
双眼线索与视觉恒常性
双眼线索要两只眼睛一起工作才能产生。视网膜差异来自左右眼看到的图像略有差别，物体越近，两幅图的差别越大，大脑用这个差值推算距离。辐合是眼肌信息，看近处物体时两眼向内转，肌肉的紧张程度被换算成距离感。视觉恒常性把这些线索的结果稳定下来，物体走远时视网膜上的像变小，我们仍然知道它的实际大小没有变。大小恒常、形状恒常与颜色恒常在日常生活里一起起作用，看一扇斜开的门不会觉得它变了形。阅卷时经常见到学生把恒常性写成错觉，两者的方向正好相反，恒常性维持真实，错觉让判断出错。
Binocular cues need both eyes: retinal disparity comes from the slightly different image each eye receives, and that difference shrinks as the object moves further away; convergence is muscle information from the eyes turning inwards to focus on something close. Visual constancies keep the world stable, so a door swinging open is still seen as rectangular instead of as a changing shape on the retina. Confusing constancy with illusion costs marks every year, because one preserves reality and the other distorts it.
错觉的四类解释
AQA 的 specification 把错觉的成因归成四类，答题时先判断题目给的是哪一类，再写解释。歧义指同一幅图有两种同样合理的读法，大脑在两者之间来回切换，Necker 立方体与 Rubin 花瓶都属于这一类。误读深度线索发生在二维图像给出虚假的远近信息时，大脑按三维世界的规则把远处的东西放大。虚构是没有轮廓的地方被看出轮廓，Kanizsa 三角形中间那个白色三角并不存在。大小恒常解释 Ames 房间，房间本身被建成不规则形状，从特定角度看却像一个正常的长方体，于是站在远角的人被判断成巨人。很多学生以为错觉说明眼睛出了故障，错觉恰恰证明知觉系统在按经验做推断。
The specification groups explanations of illusion into ambiguity, misinterpreted depth cues, fiction and size constancy, and the first job in an answer is to decide which one the stimulus illustrates. Ambiguous figures such as the Necker cube offer two equally plausible readings that the brain alternates between, while fiction describes contours seen where none exist, as in the Kanizsa triangle. The Ames room works through size constancy, since an irregular room built to look rectangular from one viewpoint makes the person in the far corner look enormous.
六个必考错觉图形
Ponzo 错觉画两条会聚的斜线，上面那条横线看起来更长，原因是大脑把它放在更远处并按比例放大。Muller-Lyer 的两条竖线本来等长，箭头向外的一条显得更长，箭头向内的显得更短。Rubin 花瓶是一图两解的经典例子，面孔与花瓶轮流出现，无法一起看到。Necker 立方体同样属于歧义，前后的面会自动翻转。Kanizsa 三角形属于虚构，三个缺口的圆让人看到一个完整的三角形。Ames 房间属于大小恒常被误用，站在远角的人看起来更高。记图形时顺手记住它对应哪一类解释，答案里就少一段犹豫。
Each named illusion in the specification is best memorised together with its explanation. The Ponzo illusion exaggerates the upper line because converging lines suggest distance, and the Muller-Lyer lines differ only in the direction of their fins; Rubin's vase and the Necker cube are ambiguous, while the Kanizsa triangle is fiction. The Ames room completes the set, using size constancy to make two people of similar height look absurdly different.
Gibson 直接理论
Gibson 认为知觉是直接的，外界提供的光学阵列已经包含全部深度与距离信息，大脑不需要推理。这个立场属于天性一侧，他主张婴儿天生具备感知三维世界的能力。运动视差是他最常用的例子，坐车时近处的树篱飞快后退，远处的山几乎不动，运动本身的差异直接给出远近。光学阵列随观察者移动而变化的方式，被 Gibson 称为不变量，它稳定地指向环境结构。按这个理论，知觉与感觉几乎是一回事，眼睛看到什么，大脑就知道什么。AQA 官网的 specification 在 3.1.2 里把运动视差单独点名，写理论时必须带上这一条。
Gibson argued that perception is direct: the optic array reaching the eye already carries enough information about distance and layout that no inference is needed, which places his theory firmly on the nature side of the debate. Motion parallax is his strongest everyday example, since hedgerows rush past a moving car while distant hills barely shift, and that difference in motion specifies depth without any guessing. In his account sensation and perception are close to the same event.
Gibson 的证据与局限
视崖实验是支持直接理论的经典证据，Gibson 与 Walk 在 1960 年让 6 到 14 个月的婴儿爬过一块中间有落差平台的玻璃，多数婴儿在深侧停住，说明深度知觉出现得很早。二战飞行员在缺少仪表时仍能准确降落，Gibson 认为他们依靠跑道给出的光学信息。这个理论的弱点同样明显，它解释不了人为什么看错，错觉里的图像信息并不缺，大脑却给出错误结论。直接理论也难以说明文化差异，同一幅画在不同成长环境里的人眼中会有不同深度。写评价题时，一条支持加一条限制，再落回天性与教养的争论，比罗列名词稳得多。
The visual cliff study by Gibson and Walk in 1960 is the standard supporting evidence, because most infants of six to fourteen months refused to crawl onto the deep side of the glass, which suggests depth perception appears early. Wartime pilots landing without instruments were said to rely on the same optic flow information as they came in over the runway. The theory struggles with illusion, where information is plentiful yet the conclusion is wrong, and it says little about cultural variation in reading pictures.
Gregory 建构理论
Gregory 把知觉看成假设检验，视觉输入是不完整的线索，大脑用过去经验补上缺失的部分，构造出一个关于外界是什么的模型。这个立场属于教养一侧，知觉是学来的，经验越多，推断越准确。他用推论这个词描述这个过程，感知一个物体等于对它是什么做出最合理的一次猜测。大小恒常在这套理论里是推断的结果，视网膜图像变小，大脑根据距离线索把物体还原成原来的尺寸。错觉因此被解释成推断出错，二维图像被当成三维场景，经验把错误的深度加了进去。AQA 官网把 Gregory 的理论排在 Gibson 之后，两条对照着记最省力。
Gregory treated perception as inference: visual input is incomplete, so the brain uses past experience to construct a hypothesis about what is out there, which makes his theory a nurture account. Size constancy illustrates the process, since a shrinking retinal image is corrected using distance cues until the object is perceived at its real size. Illusions then become errors of inference, where a flat drawing is read as a three-dimensional scene.
Gregory 的证据与局限
空心面具是常被引用的例子，面具凹进去，眼睛却把它看成凸出的脸，因为过去经验里脸总是凸的。Ames 房间同样支持建构理论，房间的真实形状与经验中的房间冲突时，大脑选择经验。Palmer 在 1975 年的实验里先让被试看厨房场景，再呈现面包与邮箱这类物体，与场景一致的东西被识别得更快。这套理论的问题在于推论过程无法直接观察，说大脑做了假设，其实是根据结果反推出来的解释。它也很难量化，不同的人经验不同，推断的强度无法测量。写评价时可以补一句神经科学的发现，视觉皮层确实存在分层处理，但把全部知觉都归给推断仍然过头。
The hollow mask illusion supports the constructivist account, because a concave mask is seen as a convex face whenever experience overrules the light actually reaching the eye; Palmer's 1975 study points the same way, with kitchen objects recognised faster after a kitchen scene. The weakness is that inference itself is inferred rather than observed, so the explanation is hard to falsify or measure. Neural evidence for staged processing in the visual cortex helps, yet treating all perception as hypothesis testing still overreaches.
文化因素与 Hudson 研究
Hudson 在 1960 年用一幅狩猎图测试深度线索的解读，图里猎人举矛、大象在远处、羚羊在前景，三个物体被安排在同一张二维画面上。南非的非洲被试与欧洲被试给出的答案差别明显，前者常常报告大象正要踩到猎人，把远处的物体读成近处。Hudson 把差异归给学校教育与读图经验，受过学校教育的被试更熟悉二维图像里的透视规则。这个研究不在 AQA 的指定名单里，官网 3.1.2 只点名 Gilchrist 与 Nesberg 和 Bruner 与 Minturn，引用时要写清它是补充证据。很多学生以为文化差异证明知觉天生不同，差异其实来自经验积累，与天生能力无关。
Hudson's 1960 picture study showed a hunting scene to African and European participants in South Africa, and unschooled viewers often read the distant elephant as close enough to trample the hunter. Hudson attributed the gap to schooling and familiarity with two-dimensional pictures rather than to any innate difference. The study is not on the AQA named list in section 3.1.2, so it works as supporting evidence and not as a core study.
动机与情绪的影响
Gilchrist 与 Nesberg 在 1952 年做的实验是 AQA 点名的动机研究，26 名大学生被分成两组，实验组连续 20 小时不进食，控制组正常吃饭。两组看同一批幻灯片，其中既有食物图片也有中性图片，任务是给每张图的亮度打分。饥饿组把食物图片评得更亮，中性图片的评分没有差别，结论是动机改变了知觉本身，而不只是改变了态度。情绪一侧常用的证据是 McGinnies 在 1949 年的研究，被试识别禁忌词比识别中性词更慢，这个现象被称为知觉防御。常见的错误是把这类研究写成情绪让人看不清，它改变的是识别的门槛与速度。
Gilchrist and Nesberg in 1952 gave 26 students slides of food and neutral scenes and asked them to rate brightness; the group deprived of food for twenty hours rated the food pictures as brighter, while neutral pictures were rated the same. Motivation therefore altered perception itself rather than a later judgement. McGinnies in 1949 reported slower recognition of taboo words than of neutral ones, an effect usually called perceptual defence.
期望 知觉定势 与答题安排
Bruner 与 Minturn 在 1955 年用一张可以看成字母 B 或数字 13 的图形做实验，先看字母的被试报告 B，先看数字的报告 13，期望由此形成知觉定势。知觉定势指先前经验让人准备好按某种方式解读当前刺激，图式一旦被激活，模糊信息就被填成预期中的形状。AQA 的 specification 把知觉定势与四个影响因素放在同一条里，答题时先写因素再写研究最清楚。Section B 只占 Paper 1 的 25 分，题型是选择题、短答和一段延伸写作，时间上给到 25 分钟就够。错得最多的是把研究和理论混着答，评价题要求同时写出支持与限制，只写一方最多拿一半分。很多学生以为多写术语更安全，阅卷时看到堆名词而没有解释的答案，通常只给一分。
Bruner and Minturn in 1955 showed an ambiguous figure that could be read as a letter B or the number 13, and the reading depended on whether participants had just seen letters or numbers, which is perceptual set created by expectation. Section B carries 25 of the 100 marks on Paper 1, so a sensible plan gives it about a quarter of the 1 hour 45 minutes. Evaluation questions want a strength and a limitation, and answers that list terms without explaining them rarely move past the first mark.
牛津高阶英汉双解词典 第10版 商务印书馆¥168📱 去京东购买京东精选
需要一对一辅导？Need One-on-One Tutoring?
A-Level / GCSE / IB / KS2-KS3 全科辅导
16621398022 同微信
Visit aleveler.com
或截图后用「扫一扫 → 相册」选择本图