高考风向预测 · 外刊精读 07Tacit Knowledge|为什么有些人明明＂会了＂,却说不清自己是怎么学会的?

Experts often know things that they find surprisingly difficult to explain. An experienced cyclist, for example, may easily keep a bicycle balanced but struggle to describe exactly how much pressure should be applied to each pedal. A skilled doctor may quickly notice something unusual in a medical image without being able to describe every mental step that led to the judgment.

This kind of hidden know-how is often called tacit knowledge. Unlike facts that can be written in a textbook or explained through a set of rules, tacit knowledge is usually developed through experience and may be difficult to put into words.

Researchers at MIT wondered whether such knowledge could be detected even when people themselves were unaware that they had learned it. If so, they asked, could this hidden knowledge eventually be made explicit and used to improve learning?

To investigate the question, the researchers recruited 30 volunteers and asked them to classify more than 120 images. Each image contained two shapes, one on each side. The shapes varied in color, form and pattern.

Unknown to the participants, only one side of each image contained information that was useful for deciding whether the image belonged to group A or group B. The other side was essentially noise.

The volunteers were not told the classification rule. At first, they had little idea how to sort the images and often had to guess. But after seeing many examples, their performance gradually improved. By the later part of the experiment, many participants were able to classify the images with much greater accuracy.

Something else had changed as well. At the beginning, participants tended to look at both sides of each image. As they became better at the task, however, their gaze increasingly shifted toward the side containing the useful information.

The surprising part was that the participants did not realize they were doing this. When the researchers asked them what strategy they had learned, the volunteers generally could not identify the relevant side of the image or explain why their performance had improved.

To examine the process more closely, the researchers used cameras to track the participants' eye movements. They also recorded brain activity using electroencephalography, or EEG.

The images were designed so that the two sides flickered at slightly different frequencies. The flickering was too subtle for participants to consciously notice, but researchers could detect which frequency was reflected in their brain activity. This provided another way to determine where a person's attention was focused, even when their eyes alone did not reveal the full story.

Both forms of measurement pointed to the same pattern. As participants became more skilled, their visual and cognitive attention increasingly favored the part of the image that actually mattered for classification.

In other words, the participants had learned something important without being able to say what they had learned. The researchers considered this a measurable form of tacit knowledge.

They then asked another question. What would happen if they made this hidden knowledge visible to the participants?

After participants had become skilled at the task, some were shown information based on their own patterns of attention. This feedback revealed where they had been unconsciously focusing when they made successful decisions.

Those participants improved their classification performance even further compared with equally trained participants who did not receive the same feedback.

The result suggests that unconscious knowledge does not necessarily have to remain unconscious. At least in some tasks, researchers may be able to detect hidden strategies and then turn them into information that a learner can deliberately use.

The findings do not mean that every kind of expertise can easily be converted into a simple rule. The study involved a controlled visual-classification task, and real-world skills are usually much more complicated.

Still, the researchers believe the approach could eventually be useful in fields where visual attention matters. These could include sports, skilled trades and medical image analysis, in which experienced people often notice important features that beginners overlook.

The broader idea is simple: people may sometimes learn more than they realize. If science can identify what experienced learners are doing differently, some knowledge that once took years to develop through trial and error might become easier to teach.

全文不要讲复杂，核心就是一条逻辑链：

这篇最典型的考点是：研究目的题、实验设计题（why did researchers do X）、研究结果题、推断题、词义猜测题、举例作用题。

文章开头先举 cycling 和 medical image 两个例子。目的不是聊自行车，也不是聊医学，而是在解释 tacit knowledge。

每张图两侧都有信息，但只有一边真正决定 Group A or Group B，另一边 just noise。这样研究人员才能观察：参与者学会以后，注意力到底有没有逐渐转向 the relevant side。如果左右两边都同样重要，就无法判断眼睛的变化是不是代表学会了隐藏规律。

MIT 研究显示，参与者刚开始时会查看图像的各个部分；随着准确率提高，他们的注意力逐步集中到研究人员预先设计的"有用"一侧。问题是，参与者自己并没有意识到这种变化。

眼动摄像机告诉研究人员 where the eyes were looking，但眼睛看哪里并不百分之百等于 where cognitive attention is focused。所以研究人员又加入 EEG：两个图形以不同、但参与者几乎感知不到的频率闪烁，研究人员可以从脑电信号里判断大脑注意力更多集中在哪一个图形上。也就是说：eye tracking 测 overt visual focus，EEG 进一步帮助测 cognitive attention。

实验后期：准确率提高，目光开始明显集中到 relevant side，EEG 也支持注意力转移。但是问参与者 What rule did you learn？很多人说不出来。三个信息放在一起——Performance improved + Attention changed + Participants couldn't explain why——研究人员认为这就是 evidence of tacit knowledge。

研究人员没有停在"We found tacit knowledge"，他们继续问：如果把参与者自己无意识使用的规律告诉他呢？于是，一些已经训练过的参与者获得了根据他们自身注意模式产生的 feedback，结果他们的表现 improved further；而同样接受过训练、但没有得到这种 feedback 的 control group，进步没有那么明显。这一步非常重要——研究不只是 detect hidden knowledge，还进一步测试 Can making it explicit improve performance？答案是，在这个实验中，yes。

科学实验文章里一看到 control group 就应该警觉。为什么不能只告诉我们"给 feedback 以后成绩提高了"？因为你还不知道成绩提高到底是不是 feedback 造成的，也可能是参与者继续练习所以自然提高。于是研究人员需要：Group A（trained + feedback）vs Group B（equally trained + no same feedback），然后比较。只有这样，才能更合理地判断 feedback 本身是不是产生了额外效果。

文章里说 experts may possess knowledge even if they cannot articulate it。上下文已经告诉你：他们会做，但是说不清。所以 articulate 在这里不是"关节"，而是 express clearly in words（清楚表达）。

文章最后没有说"以后所有技能都能这样教学"，它非常谨慎。当前实验研究的是 a controlled visual classification task，研究人员只是认为这套方法可能进一步应用于 sports、crafts、skilled trades、medical image diagnosis 等依赖视觉注意的领域。

不做长词表，这一篇留下这些：

这一篇特别适合短讲，不要逐句翻译，只抓实验设计 + 结果 + control group。