Mechanisms underlying real-time spoken language understanding

人类以每分钟180个单词的速度接触和理解口头话语。当考虑到理解一个句子必须发生的无数心理过程时，这是非常令人印象深刻的。这些包括识别单个语音，隔离单词之间的边界，从声音序列中识别单词，识别单词之间的语法关系，计算句子中单词的组合意义，以及识别说话者在情境中的交际意图。口语理解研究的一个核心目标是揭示大脑如何协调这些任务，以某种方式解释实时理解的效率，以及听众在理解语言方面的明显成功，这反映在相对较低的理解失败率上。

Humans encounter and understand spoken utterances at the rate of about 180 words per minute. This is extremely impressive when considering the numerous mental processes that must occur to understand a sentence. These include things like identifying individual speech sounds, isolating boundaries between words, recognizing words from sound sequences, identifying grammatical relationships among words, computing the combined meaning of words in a sentence, and recognizing a speaker's communicative intent in the situational context. A core goal in research on spoken language comprehension is to reveal how the mind coordinates these tasks in a way that explains the efficiency of real-time comprehension as well as listeners' apparent success at understanding language, as reflected in the comparatively low rate at which comprehension breaks down.