Do people speak faster in some languages?

Yes, but only in syllables per second. In a 2011 study that had speakers read translated versions of the same texts aloud, Japanese was the fastest of seven languages at 7.84 syllables per second, and English was slower at 6.19 (British English in the study) ^{1 2}. A words-per-minute figure for each language would need one agreed way to count words, and no study we could find gives one. So a language's words per minute tell you little about how fast it is spoken, and you should not rank languages by it.

This page explains why the answer changes with the unit, what the evidence shows, and how to measure your own pace. For the full picture of dictation languages, see the guide to dictation languages, and the languages overview for the other pages in this subject. For a related search, see the question on Dragon's languages.

Syllable rates differ between languages, but information rates do not

The 2011 study measured speakers of seven languages reading translated texts at a normal rate, and Japanese came out fastest, ahead of English.

The faster syllable rate does not mean more meaning per second. A 2019 study of 17 languages found that languages encode similar information rates, about 39 bits per second, despite large differences in speech rate and in information per syllable ³. A fast-sounding language is not necessarily carrying more meaning.

The authors of the 2011 paper explain the sound of speed with a point from Roach (1999). A language with simple syllables can fit more syllables into a second, so it "will probably sound faster" ⁴. Syllable structure can shape how fast a language sounds to a listener.

Words per minute cannot be compared across languages

A word is not a fixed unit. The same idea takes a different number of words and syllables in each language, so a word count for Japanese cannot be set against one for English. Writing adds a second problem. Mandarin Chinese has no spaces between words, and in one study native speakers agreed on where the word boundaries fall only about 75 percent of the time ⁵. A word counter has to guess where the words are, and two counters may give different totals for the same recording.

Dictation studies that report words per minute for two languages show how easily the numbers mislead. In a 2018 study of text entered on a phone, English entered by speech reached 153 words per minute, while Mandarin Chinese reached 123 words per minute ^{6 7}. Those figures measure how fast people entered text on a phone, counting the finished text after corrections, for short messages. They are not a comparison of how fast the two languages are spoken.

Speaking rates vary widely from one conversation to the next

In a study of conversational telephone speech, the speaking rate of individual conversations ranged from 111 to 291 words per minute ⁸. Each conversation was a single pair of speakers, so the range shows the spread between conversations, not between languages. If you are asking whether you speak fast enough for a task, your own speaking rate is a better measure than a figure for your language.

It depends on the kind of speech and how it is counted

The answer changes with the setting. Read-aloud speech, as in the 2011 study, is different from conversation and from a speech prepared for an audience. The 2011 figures come from read speech in a laboratory, not from spontaneous conversation. A comparison across languages is only fair when the speech type, the recording conditions and the counting method are the same for each language.

Counting also changes with the unit. A syllable count depends on the phonology of each language, and a word count depends on the spacing and tokenizing rules. For Japanese, Chinese and other languages written without spaces, a word count can differ by counter. The 2013 Mandarin study above uses character error rate as the usual metric for Mandarin speech recognition ⁵. If you see a words-per-minute figure for another language, check how the words were counted before you compare it with English.

What to do if you want to know your own pace

1. Pick one language and one counting rule. Count words the same way every time, and say which tool or method you used.
2. Time a sample of your own speech. Record five minutes of speaking in your target language, transcribe it, and count the words. Use the same counter for every sample.
3. Compare yourself with yourself. Check the same language on different days, in different settings and with different subjects. That range tells you more than any comparison with another language.
4. Do not rank languages by words per minute. If you need a cross-language comparison, look for a study that reports syllables per second and say which unit you are using.
5. For dictation, test accuracy before speed. A tool that recognizes your language poorly can cost more time than a slow speaker does. The question on which app supports the most languages explains how to compare tools, and the glossary entry on word error rate explains the accuracy measure most tests use.

For the pace of English speech, see the guide to speaking speed. The question of how accuracy varies across languages and speakers is covered in the guide to speech recognition accents, and the glossary entry on low-resource languages defines the languages that have less training data.