中文导读

窗外的鸟叫，有时听起来像有人在哼歌，这只是错觉吗？文章介绍加拿大麦吉尔大学 10 月 29 日发表在《科学》期刊上的一项研究。先交代研究规模：团队从公众分享野生动物声音的网站下载 367 种鸟类的录音，逐个人工标注音符，再与 52 种语言中的人类说话和歌唱片段比较。接着列出三种共同模式：相邻音符之间音高变化小、节奏规律、一句话或一个乐句的最后一个音往往更长。随后是最出人意料的结果：研究者原本以为像人一样靠模仿学会鸣唱的鸟会更接近人类，但证据几乎不支持这一点，天生就会鸣叫的鸟也有相似模式。作者因此推测，原因可能更多在于发声的身体条件，例如音高难以大幅跳动、句末与呼吸有关。最后，同期评论提醒，声音相似不等于意义或社会思维相同。阅读时可留意 First、Second、Third 如何分列要点，However 如何引出与预期相反的结果，以及 suggest、may、could 等表示推测的词。

正文

When a bird sings outside a window, it can sound a little like a person humming a tune. A new study suggests that this is more than a feeling. In a paper published in the journal Science on 29 October, researchers at McGill University in Canada report that bird songs and human voices share several basic sound patterns.

本段译文

当一只鸟在窗外鸣唱时，听起来可能有点像一个人在哼曲子。一项新研究表明，这不只是一种感觉。加拿大麦吉尔大学的研究人员 10 月 29 日在《科学》期刊上发表论文称，鸟类的鸣唱和人类的声音共有几种基本的声音模式。

The team, led by biologist Logan James and Professor Jon Sakata, studied 367 bird species. They downloaded recordings from xeno-canto, a website where people around the world share wildlife sounds, and then marked each note by hand. They examined more than 74,500 sound units from about 5,900 songs. They compared these with over 800 phrases of human speech and song in 52 languages from public collections.

本段译文

这个团队由生物学家 Logan James 和 Jon Sakata 教授领导，研究了 367 种鸟类。他们从 xeno-canto 下载录音——这是一个供世界各地的人分享野生动物声音的网站——然后逐个手工标注每个音符。他们分析了来自约 5,900 首鸣唱的 74,500 多个声音单元，并将其与来自公开资料库、涵盖 52 种语言的 800 多个人类说话和歌唱片段进行比较。

Three patterns appeared again and again. First, both birds and people usually move only a small distance in pitch from one note to the next. Second, their rhythms tend to be regular and easy to predict. Third, the last sound in a phrase is often longer than the sounds before it, just as a singer may hold the final note of a line.

本段译文

有三种模式反复出现。第一，鸟和人从一个音符到下一个音符时，音高通常只变化一小段距离。第二，它们的节奏往往规律且容易预测。第三，一个乐句中的最后一个声音常常比前面的声音更长，就像歌手可能会拖长一句歌词的最后一个音。

The most surprising result concerned learning. Songbirds and hummingbirds learn their songs by copying adults, much as children learn to talk. Other groups, such as suboscines and nightjars, are born knowing their songs. Because human speech and music are learned, the researchers expected learning birds to sound more like humans. However, they found very little evidence for this idea. Birds that never learn their songs showed similar patterns.

本段译文

最令人意外的结果与学习有关。鸣禽和蜂鸟通过模仿成年个体学会鸣唱，很像孩子学说话。另一些类群，比如亚鸣禽和夜鹰，生来就会它们的鸣叫。由于人类的说话和音乐是学来的，研究人员原本预计会学习的鸟听起来更像人类。然而，他们几乎没有找到支持这一想法的证据。从不学习鸣唱的鸟也表现出类似的模式。

So why do such different animals sound alike? The authors suggest that the answer may lie in the body, not only in the ability to learn. “Birds and humans are unlikely to dramatically change pitch from one note to the next because it is difficult to do,” James said. He added that longer final sounds could be linked to the way we breathe while making sounds.

本段译文

那么，如此不同的动物为什么听起来相像呢？作者们认为，答案可能在于身体，而不只在于学习能力。James 说：“鸟和人都不太可能在相邻音符之间大幅改变音高，因为这很难做到。”他补充说，句末声音更长，可能与我们发声时的呼吸方式有关。

Not every question is answered. In a separate commentary in Science, Kazuo Okanoya wrote that many features of bird and human sounds are comparable, but some comparisons still lack evidence. He noted that a key question is how far similar sounds can tell us about meaning or social thinking.

本段译文

并不是所有问题都有了答案。在《科学》上的另一篇评论中，Kazuo Okanoya 写道，鸟类和人类声音的许多特征可以比较，但有些类比仍缺乏证据。他指出，一个关键问题是：相似的声音能在多大程度上告诉我们有关意义或社会思维的信息。

Sakata hopes the work will encourage scientists to study birds that have long been overlooked: those that do not learn their songs. For now, the study offers a simple idea. When we hum a tune, we may be following some of the same physical limits as the birds outside.

本段译文

Sakata 希望这项研究能鼓励科学家去研究长期被忽视的鸟类，也就是那些不学习鸣唱的鸟。就目前而言，这项研究提供了一个简单的想法：当我们哼一段曲子时，我们可能正在遵循与窗外的鸟相同的一些身体限制。

新闻来源

• EurekAlert! (McGill University) · Birds and humans share fundamental song and speech patterns, McGill study finds (29 Oct 2026)
• Science · Pervasive patterns in bird songs resemble those in human speech and song and are independent of vocal learning (29 Oct 2026)
• Science · How comparable are bird and human sounds? (Perspective by Kazuo Okanoya, 29 Oct 2026)

✳更多文章 →