Year 13 CCEA Psychology: Oral & Listening Skills Revision | Year 13 CCEA 心理学：口语/听力备考专项

📚 Year 13 CCEA Psychology: Oral & Listening Skills Revision | Year 13 CCEA 心理学：口语/听力备考专项

This revision guide is designed for Year 13 students preparing for the CCEA A2 Psychology examination. It focuses on the psychology of spoken language – covering both speech production (oral) and speech perception (listening). The content draws from the Language and Thought option within the A2 1 module, helping you master the theories, models, key studies and evaluative points that are commonly assessed.

本复习指南专为准备 CCEA A2 心理学考试的 Year 13 学生编写，聚焦于口语心理学——既包含言语产出（口语）也包含言语知觉（听力）。内容选自 A2 1 模块中“语言与思维”选修部分，帮助你掌握常考的理论、模型、关键研究以及评价要点。

---

1. The Exam Focus on Spoken Language | 口语与听力的考试重点

In the CCEA A2 Psychology specification, spoken language topics explore how we produce and understand speech. Questions may require you to describe a model of speech production, explain errors in speaking, or discuss evidence from listening experiments. Understanding the cognitive processes behind everyday talking and hearing is essential.

在 CCEA A2 心理学大纲中，口语主题探讨我们如何产生和理解言语。考题可能要求描述言语产生的模型、解释口语中的错误，或讨论听力实验的证据。理解日常说与听背后的认知过程至关重要。

Exam tips: be ready to link theories to real-life examples, such as slips of the tongue or mishearing song lyrics. Marks are often awarded for accurate terminology and explicit use of research evidence.

考试提示：准备好将理论联系到现实生活的例子，如口误或听错歌词。评分常取决于术语的准确性和研究证据的明确使用。

---

2. Theories of Speech Production | 言语产生理论

Levelt’s (1989) model of speech production proposes four stages: conceptualisation (deciding the message), formulation (selecting words and building grammatical structure), articulation (producing sounds via the vocal tract), and self-monitoring (checking one’s own speech for errors). This framework is widely accepted because it explains the speed and fluency of normal speech.

Levelt (1989) 的言语产生模型提出四个阶段：概念化（决定要表达的信息）、言语形成（选择词汇并构建语法结构）、发音（通过声道发出声音）和自我监控（检查自己的话语是否有误）。由于该框架能解释正常言语的速度和流畅性，因此被广泛接受。

Dell’s (1986) interactive activation model offers an alternative connectionist view. Nodes representing semantic features, words and phonemes are linked, and activation can spread both top-down and bottom-up. This model accounts well for speech errors where a wrong word sounds similar to the intended one (e.g. saying “bike” instead of “bite”). However, it has been criticised for being less precise about timing.

Dell (1986) 的交互激活模型提供了另一种联结主义视角。代表语义特征、单词和音素的节点相互连接，激活既可以自上而下也可以自下而上地传播。该模型很好地解释了口误中错误词与目标词发音相似的现象（例如本想说 “bite” 却说成 “bike”）。但也有人批评它在时间精确度上不够清晰。

Model	Key features	Strength	Limitation
Levelt (1989)	Linear stages with monitoring loop	Explains self-correction; supported by reaction-time experiments	Oversimplifies interactive feedback between levels
Dell (1986)	Parallel spreading activation across semantic, lexical and phonological nodes	Excellent at predicting error types, especially phonological substitutions	Harder to test precise timing of stages

中文对照：Levelt (1989) 模型：线性阶段带监控回路；优点：解释自我修正，得到反应时实验支持；局限：过于简化了层级间的交互反馈。Dell (1986) 模型：跨越语义、词汇和语音节点的并行扩散激活；优点：擅长预测错误类型，尤其是语音替代；局限：难以检验各阶段的确切时间。

---

3. Speech Errors as a Window into the Mind | 言语错误：窥探心灵的窗口

Slips of the tongue provide valuable evidence for the units used in speech planning. According to Garrett (1975), errors can be categorised as exchanges (e.g. “lick your list” for “lick your lips”), anticipations (“a leading list” for “a reading list”) or perseverations (“beef needle” for “beef noodle”). These patterns suggest that words are planned before their individual sounds.

口误为言语计划中使用的加工单元提供了宝贵证据。根据 Garrett (1975)，错误可分为交换（如把 “lick your lips” 说成 “lick your list”）、提前（把 “a reading list” 说成 “a leading list”）或持续（把 “beef noodle” 说成 “beef needle”）。这些模式表明，单词在具体语音发出前已被整体计划。

The tip-of-the-tongue (TOT) phenomenon, studied by Brown and McNeill (1966), occurs when we know the meaning of a word but cannot fully retrieve its phonological form. Participants often recall the first letter or number of syllables correctly, supporting the idea that lexical access involves separate semantic and phonological stages.

由 Brown 和 McNeill (1966) 研究的“话在嘴边”现象，即我们知道某个词的含义却无法完全提取其语音形式。被试常能正确回忆首字母或音节数量，这支持了词汇通达包含独立语义阶段和语音阶段的观点。

For the exam, use specific error examples to evaluate the credibility of production models. Levelt’s model can explain TOT failures through faulty phonological encoding at the formulation stage, while Dell’s model explains blending errors by co-activated competing nodes.

考试中，使用具体的错误实例来评价产生模型的可信度。Levelt 模型可通过构想阶段语音编码失败解释 TOT 现象，Dell 模型则通过共同激活的竞争节点来解释混合型错误。

---

4. Stages of Speech Perception | 言语知觉的阶段

Listening to speech involves transforming a continuous acoustic signal into discrete words. The first stage is auditory analysis, where the ear encodes frequency, amplitude and duration. Next, phonetic categorisation maps sounds onto phonemes, and finally lexical retrieval accesses stored word representations. Context and top-down expectancy influence each level.

听力理解涉及将连续的声学信号转换为离散的词汇。第一阶段是听觉分析，耳朵编码频率、振幅和时长。接着，语音归类将声音映射到音素，最后词汇提取访问储存的单词表征。语境和自上而下的预期会影响每一层级。

Categorical perception demonstrates that listeners impose sharp boundaries on a continuous sound dimension. For example, a sound intermediate between /ba/ and /pa/ is consistently heard as one or the other, not as a blend. Liberman et al. (1957) first showed this with synthetic speech, highlighting the specialised nature of speech processing.

范畴知觉表明，听者在连续的声音维度上会施加清晰的边界。例如，介于 /ba/ 和 /pa/ 之间的声音会被稳定地听成其中一种，而非混合体。Liberman 等人 (1957) 首次用合成语音证明了这一现象，突显了言语加工的特殊性。

Exam application: you might be asked to draw a diagram showing the transformation from sound wave to word. Use the labels ‘acoustic signal → phoneme categorisation → lexical access’, and mention top-down influence from sentence context.

考试应用：可能要求你画出从声波到词汇的转换示意图。使用标签“声学信号 → 音素归类 → 词汇通达”，并提及从句语境而来的自上而下影响。

---

5. Models of Spoken Word Recognition | 口语词汇识别模型

The Cohort model, proposed by Marslen-Wilson and Tyler (1980), states that when a listener hears the beginning of a word, a ‘cohort’ of all possible words sharing that onset is activated. As more acoustic information arrives, candidates that no longer match are eliminated until one remains. This explains why we often recognise a word before hearing its ending.

Marslen-Wilson 和 Tyler (1980) 提出的队列模型认为，当听者听到一个词的开头，所有共享该起始音的单词会被激活形成一个“队列”。随着更多声音信息到达，不匹配的候选词被排除，直到剩下一个。这解释了为何我们常在听到词尾之前就已识别该词。

The TRACE model (McClelland & Elman, 1986) is a connectionist model with three interconnected layers: features, phonemes and words. Activation spreads both bottom-up and top-down, and units compete through lateral inhibition. TRACE handles noise and context well, but is criticised for being ‘too powerful’ and difficult to falsify.

TRACE 模型 (McClelland 和 Elman, 1986) 是一种具有三个互连层级的联结主义模型：特征层、音素层和单词层。激活既能自下而上也能自上而下地传播，单元之间通过侧抑制竞争。TRACE 能很好地处理噪声和语境，但被批评为“过于强大”而难以被证伪。

• Key study: Marslen-Wilson’s (1973) shadowing task found that people can repeat speech with a lag of as little as 150 ms, indicating very rapid word recognition.
• 关键研究： Marslen-Wilson (1973) 的跟读任务发现，人类能以短至 150 毫秒的延迟重复言语，表明词汇识别极其迅速。
• Evaluation: Cohort model struggles to explain how we recognise words beginning with ambiguous phonemes, whereas TRACE accounts for it via simultaneous bottom-up and top-down flow.
• 评价： 队列模型难以解释我们如何识别以模糊音素开头的词，而 TRACE 通过同时进行的自下而上和自上而下信息流进行了解释。

---

6. The McGurk Effect and Multisensory Integration | 麦格克效应与多感官整合

The McGurk effect (McGurk & MacDonald, 1976) demonstrates that visual information from lip movements can alter speech perception. When an audio recording of /ba/ is dubbed onto a video of someone saying /ga/, observers often perceive a different sound such as /da/. This shows that speech perception is multimodal, not purely auditory.

麦格克效应 (McGurk 和 MacDonald, 1976) 显示，来自嘴唇动作的视觉信息可以改变言语知觉。当把 /ba/ 的录音叠加到一个人说 /ga/ 的视频上，观察者常常会感知到一个不同的音，如 /da/。这表明言语知觉是多模态的，而不仅仅是听觉性的。

This finding challenges purely bottom-up auditory models and highlights the importance of top-down visual influence. The effect holds even when listeners know about the illusion, implying automated integration. For the exam, you can use the McGurk effect to evaluate the limitations of models that ignore visual inputs.

这一发现挑战了纯粹自下而上的听觉模型，凸显了自上而下的视觉影响的重要性。即使听者知道这个错觉，效应依然存在，说明整合是自动化的。在考试中，你可以用麦格克效应来评价那些忽略视觉输入的模型的局限性。

An application is in hearing-impaired individuals who rely more heavily on lip-reading. Real-world implications also include dubbing in films and speech perception in noisy environments, which you might refer to in scenario-based questions.

一个应用是听障人士更依赖唇读。现实意义还包括电影配音和嘈杂环境中的言语知觉，你可以在情境题中引用这些例子。

---

7. Listening Comprehension and Working Memory | 听力理解与工作记忆

Understanding extended speech relies on working memory, particularly the phonological loop described by Baddeley and Hitch (1974). The phonological store holds auditory input briefly, while the articulatory rehearsal process refreshes it. This system is vital for holding spoken sentences in mind long enough to extract meaning.

理解连续的话语离不开工作记忆，尤其是 Baddeley 和 Hitch (1974) 描述的语音环路。语音储存短时间内保留听觉输入信息，而发音复述过程则进行刷新。该系统对于将听到的句子保持足够长时间以提取意义至关重要。

Evidence from word-length effect shows that longer words are harder to recall in order because they take longer to rehearse, and articulatory suppression (repeating an irrelevant sound) disrupts listening comprehension. Both findings validate the role of the phonological loop in real-time language processing.

来自词长效应的证据表明，较长的单词更难按顺序回忆，因为它们需要更长的复述时间；发音抑制（重复一个无关的声音）会干扰听力理解。这些发现都证实了语音环路在实时语言加工中的作用。

In the exam, connect working memory models to listening difficulties. For instance, when a person tries to listen to instructions while counting aloud, their comprehension drops – a clear demonstration of the phonological loop’s function.

考试中，将工作记忆模型与听力困难联系起来。例如，当一个人一边大声计数一边听指令时，理解力会下降——这清晰地展示了语音环路的功能。

---

8. Factors Affecting Listening in Everyday Contexts | 影响日常听力的因素

Several psychological factors influence how well we understand spoken messages. Context and expectation speed up word recognition, as shown by the phonemic restoration effect (Warren, 1970): listeners restore missing phonemes when the missing sound is replaced by a cough, and they do not notice the interruption.

有多项心理因素会影响我们理解口语信息的程度。语境和预期会加快词汇识别速度，如音位恢复效应 (Warren, 1970) 所示：当缺漏的声音被咳嗽声替代时，听者会自动恢复缺失的音素，并且注意不到中断。

Age-related changes in hearing acuity and cognitive slowing also affect listening. Older adults may perform worse on speech-in-noise tasks. This has practical implications for designing public address systems and telephone menus. Cultural familiarity with accents can influence how easily speech is perceived.

年龄增长带来的听觉敏锐度变化和认知速度减慢也会影响听力。老年人在噪声中的言语任务表现可能更差。这对公共广播系统和电话菜单的设计有实践意义。对口音的熟悉程度也会影响言语感知的难易。

Anxiety and divided attention reduce listening efficiency. Examination context: a multiple-choice question could present a scenario of a student trying to listen to a lecture while reading social media – use theories of attention and the bottleneck model to explain the breakdown.

焦虑和注意力分散会降低听力效率。考试情境：一道多选题可能呈现一个学生一边看社交媒体一边听讲座的场景——运用注意力理论及瓶颈模型来解释理解失败的原因。

---

9. Research Methods in the Study of Spoken Language | 口语研究中的研究方法

CCEA exam questions often ask you to outline a research method used in language studies. Laboratory experiments using reaction times (e.g., shadowing tasks, lexical decision tasks) provide objective data. However, such tasks may lack ecological validity because they use isolated words rather than natural conversation.

CCEA 考题常会要求你概述语言研究中使用的一种研究方法。使用反应时的实验室实验（如跟读任务、词汇判断任务）能提供客观数据。然而，这类任务可能缺乏生态效度，因为它们使用的是孤立单词而非自然对话。

Observation and error collection are common in production research. Researchers collect slips of the tongue through diaries or recordings, then classify errors to test theoretical predictions. This method is high in ecological validity but can be subject to researcher bias in recording and classification.

观察和错误收集在产生研究中很常见。研究者通过日记或录音收集口误，然后将错误分类以检验理论预测。这种方法生态效度高，但可能因记录和分类时的研究者偏误而受影响。

Method	Strength	Weakness
Lab experiment (e.g. phoneme monitoring)	High control; replicable	Artificial stimuli; low mundane realism
Naturalistic observation of speech errors	Real-life errors; high ecological validity	Cannot establish cause and effect; potential observer bias
Case studies of language disorders	In-depth insight into broken processes	Findings may not generalise

中文对照：实验室实验：控制强，可重复；但刺激人工化，世俗现实主义低。自然观察口误：真实错误，生态效度高；不能确定因果关系，可能存在观察者偏差。语言障碍个案研究：深入了解断裂过程；但结果可能无法推广。

---

10. Key Studies to Reference | 必须引用的关键研究

Examiners expect you to name and describe specific studies. Below is a quick-reference list of essential research for oral and listening topics:

考官期望你说出并描述具体研究。以下是口语与听力主题必引用的核心研究速查列表：

• Levelt (1989): Reaction-time experiments supporting staged speech production. / 反应时实验支持分阶段言语产生。
• Dell (1986): Computational simulation of speech error patterns. / 对言语错误模式的计算模拟。
• Garrett (1975): Classification of slips of the tongue into exchanges, anticipations and perseverations. / 将口误分为交换、提前和持续等类型。
• Brown & McNeill (1966): Tip-of-the-tongue studies showing partial phonological recall. / 舌尖现象研究显示部分语音信息的回忆。
• Marslen-Wilson & Tyler (1980): Cohort model of spoken word recognition. / 口语词汇识别的队列模型。
• McClelland & Elman (1986): TRACE model testing categorical perception and context effects. / 检验范畴知觉和语境效应的 TRACE 模型。
• McGurk & MacDonald (1976): Audiovisual integration in speech perception. / 言语知觉中的视听整合。
• Baddeley & Hitch (1974): Working memory model with phonological loop. / 包含语音环路的工作记忆模型。
• Warren (1970): Phonemic restoration effect demonstrating top-down influence. / 音位恢复效应证明自上而下的影响。

---

11. Applying Theory to Exam Scenarios | 将理论应用于考试情境

CCEA A2 papers often include scenario-based questions. You might be given a short description of someone mishearing a word in a noisy café or producing a slip of the tongue during a presentation. Your job is to identify the relevant psychological concept and explain it using the correct model and terminology.

CCEA A2 试卷常包含情境题。你可能会看到一段简短描述，例如有人在嘈杂的咖啡馆听错一个词，或在演讲中出现了口误。你的任务就是找出相关的心理学概念，并用正确的模型和术语加以解释。

Example: ‘While ordering coffee, Alice heard “tea” when the barista actually said “key”.’ You could link this to the Cohort model – the initial cohort activated by /k/ included both “coffee” and “key”, and environmental noise prevented proper elimination of “tea”, a top-down influence from the context of a café adding to the error.

例如：“爱丽丝在点咖啡时，听到服务员说的是‘tea’，但人家实际说的是‘key’。”你可以联系队列模型——由 /k/ 激活的初始队列包含了 “coffee” 和 “key”，环境噪音导致未能正确排除 “tea”，咖啡馆这一语境产生的自上而下影响加剧了错误。

Practice linking each model to a unique strength and limitation. Then connect it to an everyday situation; this will improve your application marks. Use the sentence stem: ‘According to [model], this misperception occurs because…’

练习将每个模型联系到一个独特的优点和局限，再将其与一个日常情境关联起来；这会提高你的应用分。使用句式：“根据[模型]，这种误解发生的原因是……”

---

12. Revision Checklist and Final Tips | 复习清单与最终建议

Use this checklist to ensure you have covered all major topics:

使用以下清单确保所有重要主题都已复习到位：

• Can I describe and compare Levelt’s and Dell’s models? / 我能描述并比较 Levelt 和 Dell 的模型吗？
• Do I know at least three types of speech errors with examples? / 我是否知道至少三种口误类型及其例子？
• Can I explain categorical perception and the McGurk effect? / 我能解释范畴知觉和麦格克效应吗？
• Am I able to outline the Cohort model and the TRACE model, including evaluation? / 我能概述队列模型和 TRACE 模型，并包括评价吗？
• Do I understand the role of the phonological loop in listening? / 我是否理解语音环路在听力中的作用？
• Can I discuss research methods used to study speech? / 我能讨论研究言语的方法吗？
• Have I practised applying theories to short scenarios? / 我是否练习过将理论应用于短情境？

On the day of the exam, read the question carefully – look for cues like ‘speech production’, ‘slip of the tongue’, or ‘mishearing’ that signal which model to use. Manage your time well and always support your evaluation with evidence.

考试当天，仔细读题——寻找像“言语产生”、“口误”或“听错”之类的提示词，它们会提示你该使用哪个模型。合理安排时间，始终用证据支持你的评价。

Published by TutorHao | Psychology Revision Series | aleveler.com

更多咨询请联系16621398022（同微信）