How do our brains process speech? - Gareth Gaskell

翻译人员: Wanting Zhong 校对人员: Yolanda Zhang

The average 20 year old knows between 27,000 and 52,000 different words.

20 岁的普通人认识 约 27000 到 52000 个单词。

By age 60, that number averages between 35,000 and 56,000.

60 岁时，这个数字平均会达到 35000 到 56000 之间。

Spoken out loud, most of these words last less than a second.

大部分这些词讲出来 仅需不到一秒钟的时间。

So with every word, the brain has a quick decision to make:

因此每听到一个词时， 大脑都要快速做出决定：

which of those thousands of options matches the signal?

在成千上万的选项中， 到底哪个和接收到的信号相符？

About 98% of the time, the brain chooses the correct word.

大约 98% 的情况下， 大脑都能选出正确的词。

But how?

但它是怎么做到的呢？

Speech comprehension is different from reading comprehension,

言语理解与阅读理解不同，

but it’s similar to sign language comprehension—

却与手语理解相似——

though spoken word recognition has been studied more than sign language.

尽管与言语辨识相关的研究 比手语要多。

The key to our ability to understand speech

我们能够理解言语的关键

is the brain’s role as a parallel processor,

在于大脑的并行处理能力，

meaning that it can do multiple different things at the same time.

也就是说，大脑可以 在同一时间做多件不同事情。

Most theories assume that each word we know

大多数理论都假设， 我们所知道的每个词

is represented by a separate processing unit that has just one job:

均由各自的处理单元掌管， 每个单元唯一的任务是

to assess the likelihood of incoming speech matching that particular word.

评估输入语音与该单元负责的词语 相匹配的程度。

In the context of the brain, the processing unit that represents a word

在大脑中，每个单词的处理单元

is likely a pattern of firing activity across a group of neurons

可能就是大脑皮层中的

一群神经元的放电模式。

in the brain’s cortex.

When we hear the beginning of a word,

当我们听到词语的开头时，

several thousand such units may become active,

可能会有上千个这样的单元被激活，

because with just the beginning of a word,

因为仅知道词语开头的话，

there are many possible matches.

是有很多种匹配可能的。

Then, as the word goes on, more and more units register

随着单词的继续， 越来越多的单元发觉

that some vital piece of information is missing and lose activity.

它们缺失了一些关键信息， 随即不再活跃。

Possibly well before the end of the word,

可能早在单词结束之前，

just one firing pattern remains active, corresponding to one word.

就只剩一种放电模式仍旧活跃， 对应一个单词。

This is called the "recognition point."

这叫做 “识别点”。

In the process of honing in on one word,

在逐渐筛选出一个词的过程中，

the active units suppress the activity of others,

活跃的单元会抑制 其他单元的活动，

saving vital milliseconds.

节省关键的几毫秒时间。

Most people can comprehend up to about 8 syllables per second.

大多数人每秒钟能理解 多达 8 个音节。

Yet, the goal is not only to recognize the word,

但目标不仅仅是识别出单词，

but also to access its stored meaning.

还要提取它储存的意思。

The brain accesses many possible meanings at the same time,

在完全辨认出单词之前，

大脑已同时提取出 很多可能的词义。

before the word has been fully identified.

We know this from studies which show that even upon hearing a word fragment—

这是因为有研究显示， 即使只听到一个单词片段——

like "cap"—

比如 “cap”——

listeners will start to register multiple possible meanings,

听者都会开始想到 若干可能的词义，

like captain or capital, before the full word emerges.

比如 “captain” 或 “capital”， 即使完整的单词还没有显现。

This suggests that every time we hear a word

这表示每次听到一个单词时，

there’s a brief explosion of meanings in our minds,

我们的脑海中都会短暂地 迸发出许多涵义，

and by the recognition point the brain has settled on one interpretation.

而等到了识别点时， 大脑已经确认了一种释义。

The recognition process moves more rapidly

比起一连串随机词汇，

with a sentence that gives us context than in a random string of words.

识别过程能更迅速地辨识 提供了语境的句子。

Context also helps guide us towards the intended meaning of words

语境也能引导我们 正确地理解多义词，

with multiple interpretations, like "bat," or "crane,"

比如 “bat（蝙蝠/球棒）” 或 “crane（鹤/起重机）”，

or in cases of homophones like "no" or "know."

或者同音字，比如 “no（不）” 和 “know（知道）”。

For multilingual people, the language they are listening to is another cue,

对于掌握多种语言的人来说， 他们在听的语言也是一种线索，

used to eliminate potential words that don’t match the language context.

能用来排除 并不符合语言情景的待选单词。

So, what about adding completely new words to this system?

那么把完全陌生的词 加到这个系统中会怎样呢？

Even as adults, we may come across a new word every few days.

即使作为成人，我们每几天 也可能遇到一个新词。

But if every word is represented as a fine-tuned pattern of activity

但如果每个词都由 分散在许多神经元中的

distributed over many neurons,

精准的激活模式所控制，

how do we prevent new words from overwriting old ones?

我们该如何避免新词覆盖旧词？

We think that to avoid this problem,

我们认为，为了避免这个问题，

new words are initially stored in a part of the brain called the hippocampus,

新词最初是储存在大脑中 一个叫做 “海马体” 的区域，

well away from the main store of words in the cortex,

远离大脑皮层中的词汇主存储器，

so they don’t share neurons with others words.

因此它们不会和其他单词 共用神经元。

Then, over multiple nights of sleep,

然后，经过几晚睡眠，

the new words gradually transfer over and interweave with old ones.

这些新词会逐渐转移过去， 和旧词交织在一起。

Researchers think this gradual acquisition process

研究者们认为， 这种渐进的习得过程

helps avoid disrupting existing words.

能帮助避免扰乱现存的词汇。

So in the daytime,

所以在白天，

unconscious activity generates explosions of meaning as we chat away.

当我们聊天时，无意识的活动 制造出了词义的礼花；

At night, we rest, but our brains are busy integrating new knowledge

到了晚上，我们进入了睡眠， 而大脑则会开始忙着把新知识

into the word network.

整合进词汇网络中。

When we wake up, this process ensures that we’re ready

在我们醒来时，这个过程能确保我们

为瞬息万变的语言世界做好了准备。

for the ever-changing world of language.