We’re Only Beginning To Understand Dogs’ Facial Expressions#

狗狗的表情代表什么？你可能真的不懂#

文章导读
人类天生会读脸——连别的物种的脸也读。狗的脸有表情吗？有，而且可能比你以为的更难读懂。
文章的骨架是"工具 → 案例 → 反转 → 人群差异 → 可靠结论"：先介绍 1978 年为人类表情开发的 FACS 编码系统衍生出的 Dog FACS（20 多个价值中性的面部动作描述符）；再以最出名的"小狗眼"为例——皱眉露眼白的能力是狼没有的，可能是狗在驯化中"钻了人类喜爱幼态特征的空子"，收容所里会做这表情的狗被重新领养得更快；随即反转：那副"内疚相"其实是害怕主人不高兴，不是内疚。
后半部分转向"人读得怎么样"：婴儿反而比成人更准，养狗多年的人更倾向于把所有叫声当敌意，四岁儿童近 70% 把呲牙狗看成开心——最后给出三条可靠的归纳（耳朵、牙齿、眼睛），并收在一句话上：光看脸不够，要结合整体行为。

阅读提示：文中加粗的词为四级 / 六级 / 考研高频词，文末附核心词汇表与长难句解析。个别专业词（如 neotenic、sclera）已单独列出并注明"超纲"，不必强记。

---

一、正文精读（英中对照）#

第 1 段：人是天生的"读脸机器"#

Humans are hardwired to read facial expressions because they’re integral to social communication in our species. Our brains even scan the mugs of other species to discern their mood and intent. And most people who have interacted with dogs can agree they too have expressive faces, to one degree or another.

译文：人类天生就会读面部表情，因为表情在我们这个物种的社会沟通中不可或缺。我们的大脑甚至会去端详其他物种的脸，来辨别它们的情绪与意图。而大多数与狗打过交道的人都同意：狗的脸也在不同程度上富于表情。

翻译要点

• hardwired 原为电路术语"硬接线的"，引申为"天生的、刻在本能里的"，be hardwired to do 即"天生就会做……"，是科普文本高频词。
• integral to 意为"对……不可或缺的"，是考研高频搭配，与后文 central to 近义。
• mugs 在此为口语用法，指"脸、面孔"（略带诙谐），不是"马克杯"——熟词僻义。
• discern 意为"辨别、看出（需要仔细分辨的东西）“，比 see 强调"费心分辨”。
• to one degree or another 意为"或多或少、在不同程度上"。

---

第 2 段：分析工具——从 FACS 到 Dog FACS#

In fact, some of the most detailed analyses of how animals use their faces are derived from the Facial Action Coding System (FACS) developed to characterize human expressions in 1978. The Dog FACS has identified over 20 descriptors of canine facial movements. These descriptions, however, are value-neutral — they don’t ascribe particular emotions to the movements. A number of studies, however, have attempted to do so.

译文：事实上，关于动物如何使用面部的一些最细致的分析，都源自 1978 年为刻画人类表情而开发的面部动作编码系统（FACS）。犬类 FACS 已识别出 20 多个犬类面部动作的描述符。不过，这些描述是价值中立的——它们并不把特定情绪归给这些动作。然而，已有不少研究尝试这样做。

翻译要点

• derive from 意为"源自、衍生自"，学术写作高频；be derived from 为被动形式。
• characterize 意为"描述……的特征、刻画"，是研究方法类文本的关键动词。
• descriptor 指"描述符、描述性标签"，编码系统的基本单位。
• canine 为形容词"犬的、狗的"，正式用语（名词可指"犬齿"）。
• value-neutral 意为"价值中立的"，即"只描述动作、不判断情绪"——这是后面所有研究的科学前提。
• ascribe A to B 意为"把 A 归给 B"，与 attribute A to B 同义，考研高频。

---

第 3 段：小狗眼——最出名的表情#

Perhaps the most recognizable dog expression: the famous “puppy dog eyes.” In fact, dogs’ ability to furrow their brows is uniquely appealing to humans: this feature may have played an important role in the domestication process.

译文：也许最容易被认出的狗的表情：著名的"小狗眼"。事实上，狗皱眉的能力对人类有独特的吸引力：这一特征可能在驯化过程中发挥了重要作用。

翻译要点

• furrow 本义"犁沟"，作动词指"起皱"，furrow one's brows 即"皱眉"。
• appealing to 意为"对……有吸引力的"，appeal to 的形容词化用法。
• domestication 意为"驯化、驯养"，源自 domesticate，动物行为学核心词。

---

第 4 段：驯化假说——钻了"幼态偏爱"的空子#

One hypothesis suggests that dogs’ early ancestors were wolves who became accustomed to human presence due to a dependency on their food waste. They made the jump to full partners with their hominid commensals by hacking our attraction to neotenic features — humans are innately drawn to infantile characteristics like big eyes and short snouts.

译文：一种假说认为，狗的早期祖先是狼——因为依赖人类丢弃的食物残渣，它们逐渐习惯了人类的Presence。它们通过"钻"人类对幼态特征的偏好这一空子，一跃成为与远古人类共生的完整伙伴——人类天生就被婴儿般的特征吸引，比如大眼睛和短口鼻。

翻译要点

• hypothesis 意为"假说"，复数 hypotheses，科研文本提出解释时的标准用词。
• be accustomed to 意为"习惯于"，to 为介词。
• commensal 原为生物学术语"共生物、食客"，指一种生物栖居于另一种生物身边并获益；此处幽默地指"蹭吃蹭住的远古人类邻居"，属超纲词，理解即可。
• hack 本义"非法侵入"，此处引申为"钻……的空子、利用（系统漏洞）"——把人类的幼态偏爱当成可以被"攻破"的系统，是生动的拟计算机比喻。
• neotenic 意为"幼态延续的"（成年个体保留幼体特征），专业词，超纲。
• innately 意为"天生地"，与第 1 段 hardwired 同义呼应。
• infantile 意为"婴儿般的、幼儿的"。

---

第 5 段：皱眉露眼白的进化账#

The ability to wrinkle the brow — which is not found in wolves — exposes the sclera, or white of the eye, and is one of a number of factors that gives domestic dogs a more puppy-like appearance. An analysis of shelter adoptions found that dogs who make this expression are rehomed more quickly, reinforcing its adaptive benefit.

译文：皱眉的能力——狼并没有——会让眼白（巩膜）露出来，这是让家犬显得更像幼崽的诸多因素之一。一项对收容所领养数据的分析发现，会做这种表情的狗被重新领养的速度更快，这进一步印证了它的适应性收益。

翻译要点

• wrinkle 作动词"使起皱"，wrinkle the brow 即"皱起眉头"。
• exposes the sclera, or white of the eye 中 or white of the eye 是 sclera 的通俗同位解释——专业词后立刻给通俗说法，是科普文的标志手法。sclera（巩膜）超纲。
• one of a number of factors that gives... 中 that gives 的单数动词以 one 为逻辑主语。
• shelter adoption 指"（动物）收容所的领养"，shelter 在此不是"庇护所"的抽象义，而是"流浪动物收容所"。
• rehome 为"re- + home"构词动词，“为……重新找到家”，即"被再次领养"。
• reinforcing its adaptive benefit 为现在分词短语作结果状语，adaptive 意为"适应性的"。

---

第 6 段：反转——"内疚相"其实是害怕#

We may misread this look, though. Another study found that dogs who ate a forbidden treat made puppy dog eyes and were perceived as looking “guilty” after they were scolded for the transgression. But the dogs in the study made the same “guilty” look when they were scolded and hadn’t done anything wrong. It seems that the expression was more a result of distress at their handlers’ disapproval than any true feelings of guilt.

译文：不过，我们可能会误读这种表情。另一项研究发现，偷吃了禁食零食的狗会摆出小狗眼，在被主人为此过错训斥之后，被感知为看起来"很内疚"。但研究中那些被训斥了、却什么坏事也没做的狗，摆出的也是同样的"内疚相"。看来，这种表情更多是出于对饲养者不悦的惶恐，而不是任何真正的内疚感。

翻译要点

• misread 为"mis- + read"构词动词"误读"，与标题 We're Only Beginning To Understand 的谨慎立场呼应。
• forbidden treat 意为"被禁止吃的零食"，forbidden 为过去分词作定语。
• be perceived as doing 意为"被感知为、被认为"，perceive 是考研高频动词。
• scold 意为"训斥、责骂"，be scolded for sth. 即"因某事被骂"。
• transgression 意为"违规、过失"，比 mistake 更正式，此处带幽默感。
• 末句的 more A than B 结构是实验结论的标准表述：“与其说是 B，不如说是 A”。
• distress 意为"痛苦、惶恐不安"，与 guilt（内疚）形成科学区分——狗怕的是主人不高兴，不是良心发现。

---

第 7 段：人类读狗脸——婴儿反而更准#

Understanding some basic dog expressions appears to be instinctive on some level. An experiment that exposed human infants to images of dogs exhibiting both playful and aggressive expressions found that they were able to accurately match them to correlating barks.

译文：在某种程度上，理解狗的一些基本表情似乎是本能。一项实验让人类婴儿观看既有玩耍表情、也有攻击表情的狗的图片，发现他们能够准确地把这些表情与相应的叫声匹配起来。

翻译要点

• instinctive 意为"本能的"，与第 1 段 hardwired、第 4 段 innately 构成同义链。
• expose A to B 意为"使 A 接触 B"，实验设计术语（对照组/实验组设计常用）。
• exhibiting 为现在分词短语作后置定语修饰 dogs。
• correlating barks 中 correlate 意为"相关、对应"，correlating 即"相应的、对应的"。

---

第 8 段：成人反而更差#

“Adults and older kids, however, are far less accurate,” notes lead author Ross Flom, an assistant professor of psychology at Southern Utah University. “That is, older adults tend to treat all barks initially as aggressive.”

译文：""然而，成年人和大一些的孩子准确度差得多，"主要作者、南犹他大学心理学助理教授罗斯·弗洛姆指出，“也就是说，成年人起初倾向于把所有叫声都当作攻击性的。”

翻译要点

• notes lead author Ross Flom 为倒装的引语插入：正常语序是 Ross Flom notes...，把 notes + 身份 提到引语后是新闻与学术写作的常见变体。
• lead author 意为"第一作者、主要作者"，学术署名术语；assistant professor 为"助理教授"（职称，不是"助教"）。
• That is 意为"也就是说"，用于复述前句，是解释性衔接词。

---

第 9 段：经验假说——越养越偏#

He hypothesizes that this loss of ability may be due to the cumulative time spent with dogs. “Our experience modifies in such a way that we do tend to treat each bark as aggressive.” As we grow older, we generalize barks as hostile regardless of the expression on the dog’s face. Infants, however, “use acoustic subtleties to determine or to differentiate between play and aggressive barks,” he suggests.

译文：他提出假说：这种能力的丧失，可能与人和狗相处累积的时间有关。"我们的经验以这样的方式被塑造：我们确实倾向于把每一声叫都当作攻击性。"随着年龄增长，我们会把狗叫一概泛化为敌意，而不管狗脸上的表情。但婴儿"会利用声音的细微差别来判断、或区分玩耍性叫声与攻击性叫声，"他说。

翻译要点

• hypothesize 为 hypothesis 的动词形式"提出假说"，与第 4 段名词呼应。
• cumulative 意为"累积的"，the cumulative time spent with dogs 中 spent 为过去分词作后置定语。
• in such a way that... 意为"以这样一种方式，以至于……"，正式的结果状语结构。
• generalize A as B 意为"把 A 一概归为 B"，generalize 是科研文本高频词。
• regardless of 意为"不管、不顾"，介词短语。
• acoustic subtleties 意为"声音的细微差异"，subtlety 源自 subtle（微妙的）。
• differentiate between A and B 意为"区分 A 与 B"，与 distinguish A from B 同义。

---

第 10 段：爱狗的人反而更读不准#

Strangely, another investigation found that adults who were experienced with dogs were less capable of interpreting aggressive expressions — perhaps resulting from generally positive experiences among dog lovers.

译文：奇怪的是，另一项调查发现，与狗打交道经验丰富的人，反而更不善于解读攻击性表情——原因可能是爱狗人士的经历总体上是愉快的。

翻译要点

• be capable of doing 意为"有能力做某事"，less capable of interpreting 即"更不善于解读"。
• resulting from 意为"由……导致"，现在分词短语作非限定性补充。
• dog lovers 意为"爱狗人士"，generally positive experiences 暗示"正经历太多，就把呲牙也当友好"。

---

第 11 段：儿童的危险误读#

Other research casts further doubt on our innate capabilities. Younger children tend to interpret dogs’ hostile expressions as friendly: Nearly 70% of 4-year-olds viewed snarling dogs as happy, as reported in a 2011 study. This could explain dog attacks on young children who are not familiar with how to interpret their behavior.

译文：其他研究进一步质疑了我们天生的能力。年龄较小的儿童倾向于把狗的敌意表情解读为友好：2011 年的一项研究报告称，近 70% 的四岁儿童把呲牙低吼的狗看成开心的狗。这或许可以解释一些狗袭击幼童的事件——这些孩子还不熟悉如何解读狗的行为。

翻译要点

• cast doubt on 为固定搭配"使人对……产生怀疑"，cast further doubt on 即"进一步质疑"。
• innate 意为"天生的"，与 hardwired、instinctive 构成全文第三个同义词。
• snarl 意为"（动物）呲牙低吼"，snarling dogs 即"呲牙的狗"。
• as reported in a 2011 study 为"as + 过去分词"结构，指"如 2011 年一项研究所报告的"。
• 末句把实验数据与现实后果（儿童被狗咬）挂钩，是科普文"研究意义"的标准写法。

---

第 12 段：个体差异——共情程度#

Personality factors further complicate our interpretations of dog faces, just as they do for human faces. People who were rated higher in empathy, for example, attributed greater negativity to threatening dog faces than those who had lower empathy scores.

译文：人格因素让我们对狗脸的解读更加复杂，正如它们对人类面孔所起的作用一样。例如，在共情程度上得分更高的人，比共情得分低的人，把更多负面性归给了带有威胁性的狗脸。

翻译要点

• just as they do for human faces 中 as 引导比较状语从句，do 替代 complicate，避免重复。
• empathy 意为"共情、同理心"，是心理学与教育学的核心词。
• attribute A to B 意为"把 A 归给 B"，与第 2 段 ascribe A to B 同义替换——同一篇里换词，是学术写作避免重复的标准操作。
• negativity 意为"负面性、消极程度"，由形容词 negative 名词化而来。

---

第 13 段：可靠的归纳——耳朵、牙齿、眼睛#

A handful of solid generalizations have emerged, however, allowing us to refine our intuitive perceptions. A playful attitude has been inferred from the erection of the ears, exposure of the teeth, and partial closure of the eyes. Conversely, aggression is usually signified through moving the ears forward, using direct eye contact, and opening the mouth to expose the teeth. Submission is often demonstrated by flattening the ears, lowering the head, and retracting the lips.

译文：不过，一些可靠的归纳已经出现，让我们得以修正直觉上的判断。玩耍姿态可以从竖起的耳朵、露出的牙齿和半眯的眼睛推断出来。相反，攻击性通常表现为耳朵前倾、直视对方、张嘴露牙。顺从则常常表现为耳朵放平、头部放低、嘴唇后缩。

翻译要点

• a handful of 意为"少数几个、一小批"。
• generalization 意为"归纳出的通则"，与第 9 段动词 generalize 同源。
• refine 意为"细化、改进"，refine our intuitive perceptions 即"修正直觉判断"。
• infer A from B 意为"从 B 推断出 A"，此处用被动 has been inferred from。
• erection 在此指"竖起"（耳朵竖立），由动词 erect 而来，注意不是" erection "的其他引申义。
• signify 意为"表明、意味着"，比 show 更正式。
• submission 意为"顺从、屈服"，动物行为学术语，与 aggression 相对。
• retract 意为"缩回、收回"，retracting the lips 即"嘴唇后缩（抿嘴）"。
• 三个行为学描述都用"名词化 + of"结构（erection of the ears / exposure of the teeth / closure of the eyes），译文转为"动词 + 宾语"更自然。

---

第 14 段（结尾）：光看脸不够#

Generally, we appear best at interpreting our dogs’ emotional states when their whole range of behaviors, such as postural changes and vocalizations, are taken into account. So facial expressions alone don’t cut it.

译文：总体而言，当把狗的全套行为——比如姿势变化和叫声——都纳入考虑时，我们解读狗的情绪状态才最准确。所以，光看面部表情是不够的。

翻译要点

• when their whole range of behaviors ... are taken into account 中 take into account 为固定搭配"把……纳入考虑"，被动形式 are taken into account。
• 主谓一致：behaviors（复数）与 are taken 对应——尽管 their whole range of... 看起来像单数中心词，这里按复数处理（强调多种行为）。
• postural changes 意为"姿势变化"，vocalizations 意为"发声、叫声"，动物行为学标准术语。
• don't cut it 为口语习语"不够格、达不到要求"，源自 cut the mustard。全文最后一句从学术腔切换到口语，制造轻松的收尾——这正是科普文章的常用结尾法。

---

二、核心词汇表#

1. 考研 / 六级进阶词#

2. 四级高频词#

3. 专业 / 超纲词（理解即可，不必强记）#

4. 高频短语 / 固定搭配#

---

三、长难句解析#

1. The ability to wrinkle the brow — which is not found in wolves — exposes the sclera, or white of the eye, and is one of a number of factors that gives domestic dogs a more puppy-like appearance.

• 主干：The ability ... exposes the sclera ..., and is one of a number of factors ...，两个并列谓语 exposes 与 is 共用主语 The ability。
• 插入成分：— which is not found in wolves — 为非限制性定语从句，which 指 the ability to wrinkle the brow 这一整件事。
• 同位解释：or white of the eye 是 sclera 的通俗说法，or 在此意为"也就是"。
• 嵌套定语从句：that gives domestic dogs a more puppy-like appearance 修饰 factors，内含 give sb. sth. 双宾结构。
• 译文：皱眉的能力——狼并没有——会让眼白（巩膜）露出来，这是让家犬显得更像幼崽的诸多因素之一。

2. Another study found that dogs who ate a forbidden treat made puppy dog eyes and were perceived as looking “guilty” after they were called out for the transgression.

• 主干：Another study found that ...，that 引导宾语从句。
• 从句主语：dogs who ate a forbidden treat，who 从句修饰 dogs。
• 从句谓语：两个并列——made puppy dog eyes 与 were perceived as looking "guilty"（被动）。
• 时间状语从句：after they were called out for the transgression，call out 此处意为"点名批评"。
• 译文：另一项研究发现，偷吃了禁食零食的狗会摆出小狗眼，在被主人为此过错训斥之后，被感知为看起来"很内疚"。

3. Generally, we appear best at interpreting our dogs’ emotional states when their whole range of behaviors, such as postural changes and vocalizations, are taken into account.

• 主干：we appear best at interpreting ...，appear best at doing 意为"在做……时表现最好"。
• 时间状语从句：when their whole range of behaviors ... are taken into account，被动语态。
• 插入例举：such as postural changes and vocalizations 用两个逗号隔开，举例说明 behaviors。
• 主谓一致：中心词 behaviors 为复数，故从句谓语用 are taken。
• 译文：总体而言，当把狗的全套行为——比如姿势变化和叫声——都纳入考虑时，我们解读狗的情绪状态才最准确。

---

四、写作素材提炼#

可迁移的观点句

1. Humans are hardwired to read facial expressions because they're integral to social communication in our species.（hardwired + because 因果开场，适合"本能 vs 学习"类话题）
2. These descriptions, however, are value-neutral — they don't ascribe particular emotions to the movements.（先立科学前提，再进入研究，体现"描述"与"解释"的区分）
3. It seems that the expression was more a result of distress at their handlers' disapproval than any true feelings of guilt.（more A than B 审慎结论句式，纠正大众误读的标准写法）
4. So facial expressions alone don't cut it.（口语化收尾，一锤定音地否定单一证据来源）

可迁移的论证结构

工具（FACS → Dog FACS，价值中立）→ 明星案例（小狗眼）→ 进化解释（幼态假说）→ 数据支持（收容所领养）→ 反转（"内疚"实为惶恐）→ 转向人类读脸能力（婴儿 vs 成人）→ 人群差异（爱狗者 / 四岁儿童 / 共情者）→ 可靠归纳（耳、齿、眼三态）→ 收束（光看脸不够）

注意第 5 步的反转写法：作者先用收容所数据确立"小狗眼有用"，再用对照实验拆穿"内疚"的误读——同一表情，两种解释，先扬后抑，是纠正流行误解的标准动作。

数据引用句式

• Nearly 70% of 4-year-olds viewed snarling dogs as happy, as reported in a 2011 study.（数据 + as reported 出处标注）
• An analysis of shelter adoptions found that dogs who make this expression are rehomed more quickly.（分析来源 + 发现）

可迁移的术语块

• Facial Action Coding System (FACS) 面部动作编码系统 ｜ puppy dog eyes 小狗眼 ｜ neotenic features 幼态特征 ｜ intrusive thoughts 侵入性思维（对比参考） ｜ take into account 纳入考虑

---

五、一句话总结#

狗脸确实有表情，但人读它主要靠经验而非本能：婴儿比成人准，养狗的反而更容易把一切当敌意——而那副著名的"内疚相"，只是怕你生气，不是认错。

---

本文档由截图原文整理，英文原文版权归原作者及扇贝阅读所有，译文与词汇标注仅供学习使用。