How do our brains process speech? - Gareth Gaskell

譯者: Ann Chen 審譯者: Pui-Ching Siu

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

一般 20 歲的人知道 27,000 至 52,000 個不同的字。

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

60 歲前，該數目一般在 35,000 至 56,000 之間。

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

如果將它們唸出來， 大部份的字不用花到一秒鐘。

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

聽到每個字時， 大腦都必須快速做決定：

which of those thousands of options matches the signal?

上千個選項中的哪一個 吻合你聽到的聲音？

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

大約 98% 的時間， 大腦會選出正確的字。

But how?

但大腦如何做到呢？

Speech comprehension is different from reading comprehension,

語音理解不同於閱讀理解，

but it’s similar to sign language comprehension—

但類似於手語理解──

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

雖然人們對口語辨識的 研究比手語多。

The key to our ability to understand speech

我們能夠理解語音的關鍵

is the brain’s role as a parallel processor,

在於大腦能夠並行處理，

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

意思是它可以同時做多種不同的事。

Most theories assume that each word we know

大多數理論假設 我們所知道的每個字

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

分別由只負責一項工作的 處理單元來代表：

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

該工作是評估傳入語音 與該特定單字匹配的可能性。

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

對大腦來說，代表單字的處理單元

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

很可能是大腦皮層中 一組神經元的激發活動。

in the brain’s cortex.

When we hear the beginning of a word,

當聽到一個字的開頭音節時，

several thousand such units may become active,

可能會有幾千個處理單位活躍起來，

because with just the beginning of a word,

因為當只有字的開頭音節，

there are many possible matches.

會有很多可能的匹配。

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

繼續講字的其他音節時，

更多的單元會發現 缺乏某個重要的訊息，

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

隨之失去了活性。

Possibly well before the end of the word,

可能在字的結尾前，

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

只有一個激活模式保持活躍， 並對應於一個字，

This is called the "recognition point."

這就是所謂的「識別點」。

In the process of honing in on one word,

在琢磨一個字的過程中，

the active units suppress the activity of others,

活躍的單位會抑制其他單位的活動，

saving vital milliseconds.

從而省下重要的毫秒。

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

大多數人每秒最多可理解 8 個音節。

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

可是目標不只是辨識那個字，

but also to access its stored meaning.

還要找出內在含義。

The brain accesses many possible meanings at the same time,

在字還未完全辨識出之前，

before the word has been fully identified.

大腦會同時找出許多可能的字義。

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

我們從研究中瞭解到， 即使一聽到字的片段──

like "cap"—

譬如 「首」──

listeners will start to register multiple possible meanings,

聽者也會開始找多個可能的字義，

like captain or capital, before the full word emerges.

比如「首領」或「首都」， 在整個字唸完前就會這麼做。

This suggests that every time we hear a word

這間接說明我們每聽到一個字時，

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

腦中會很快地迸出許多字義，

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

到達辨識點時， 大腦已選中一個字義。

The recognition process moves more rapidly

如果提供有上下文的句子，

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

而不是隨機的一串單字時， 識別過程的速度會加快。

Context also helps guide us towards the intended meaning of words

上下文也會導引我們瞭解 具有多種解釋的字的正確意義。

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

例如「蝙蝠（球棒）」 或「吊車（鶴）」，

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

或在同音字的情況下， 例如 “no” 或是 “know”。

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

對多種語言的人而言， 他們正在聽的語言是另一種提示，

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

能夠去除不符合這語言的可能單字。

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

那麼，如果要在此系統中 加入全新的字呢？

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

即使是成年人，也可能每隔幾天 就會遇到一個新的字。

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

但是，如果每個字都視為 一種精細活動模式，

distributed over many neurons,

且分佈在許多神經元上，

how do we prevent new words from overwriting old ones?

我們如何避免新的字覆蓋舊的字呢？

We think that to avoid this problem,

我們認為，為避免這個問題，

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

新的字會先存在大腦海馬區中，

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

離大腦皮層中的 主要單字存儲區很遠，

so they don’t share neurons with others words.

所以它們不與其他單字共用神經元。

Then, over multiple nights of sleep,

經由多個夜晚的睡眠，

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

新字逐漸轉移過來， 與舊字交織在一起。

Researchers think this gradual acquisition process

研究人員認為， 這種緩慢的新字獲得過程

helps avoid disrupting existing words.

有助於避免破壞現有單字。

So in the daytime,

所以在白天聊天時，

unconscious activity generates explosions of meaning as we chat away.

無意識的神經元活動 會迸出許多字義來；

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

到了晚上，我們休息，

可是大腦忙著將新知識 整合到單字網絡中。

into the word network.

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

這個過程確保我們醒來時，

for the ever-changing world of language.

已經為不斷變化的 語言世界做好準備。