Rate of Speech and Comms Efficiency

RATE OF SPEECH AND COMMUNICATION EFFICIENCY FOR DIFFERENT LANGUAGES

  • The rate of speech (in syllables per second, SPS) across languages is an important and underreported topic. Robb et al (2003) found differences in speech rate for varieties within a language. New Zealand English was produced at a faster rate than American English. However, it appears that only Pelligrino and colleagues (2004, 2007, 2011) have reported different rates of speech for different languages.  
  • Click to download Pelligrino text

    In our study, we also found differences in rate of speech for American English (AE), Arabic, and Tamil (Wilson & Netsell, 2009).  Tamil was fastest (6.5 SPS), AE was the slowest (3.5 SPS), and Arabic was intermediate (5.5 SPS).*

    Click to download Poster

    Even though a given language might be spoken faster than another, it may be the information content is essentially the same. For example, the word “heat” is one syllable in AE, and four syllables in Spanish (“caliente”).  In a post-hoc analysis of our data, we computed a “Communication Efficiency Ratio” (CER). CER was calculated as Syllables per Second / Words per Second. As shown below, the CERs for AE, Arabic, and Tamil were 3.0, 2.6, and 2.5, respectively.  These CER values were not statistically significant (Chi-square comparisons).

     

    Communication Efficiency Ratio.  Syllables per second (SPS), syllables per word (SPW),

    and CER (Communication Efficiency Ratio SPS:SPW) for all participants and languages. 

    M = Mean. P = participant number.

    _________________________________________________________________________

     AMERICAN ENGLISH                        ARABIC                            TAMIL

           SPS   SPW CER               SPS  SPW  CER              SPS  SPW  CER  

    P1    3.4    1.3     2.6               P1   6.4   2.2    2.9          P1    6.7     2.8     2.4

    P2    4.4    1.2    3.7               P2    6.0   2.5     2.4          P2    5.5     2.4     2.3

    P3    3.4    1.3    2.6               P3    5.6   2.2    2.5          P3    6.5     2.4     2.7

     

    M      3.7   1.3      3.0                       6.0   2.3    2.6                  6.2     2.5     2.5

    ________________________________________________________________________

     

    In summary, from this limited sample, we conclude these languages are spoken at different rates, but their communication efficiency is the same. It seems unlikely that languages are different in terms of communication efficiency.  Pelligrino et al reached the same conclusion using a different index, i.e. “Information Density.”

    *NOTE: The above rates in our study were counted by the first author. Subsequently, a native Arabic speaker did a separate syllable count, and found there was no difference in the rates for Tamil and Arabic.  This underscores the importance of having native speakers compute SPS for their language.  The values in the above table are from the Arabic speaker. We thank Fadi Najem for these data.

    REFERENCES

    Pellegrino, F., Farinas, J., & Rouas, J-L (2004). Automatic Estimation of Speaking Rate in Multilingual Spontaneous Speech.  Speech Prosody.  Nara, Japan

    http://www.isca-speech.org/archive_open/sp2004/sp04_517.pdf

    Pellegrino, F, Coupe, C., & Marisco, (2007). An information theory-based approach to the balance of complexity between phonetics, phonology and morphosyntax. Paper presented at the annual meeting of the Linguistic Society of America, Anaheim, CA.

    Pelligrino et al, 2011.  A cross-language perspective on speech. Language 87(3): 539-558).

    Robb, M., Maclagan, M., & Chen, Y. (2003). Speaking rates of American and New

    Zealand varieties of English.  Clinical Linguistics and Phonetics 17: 1-15.  

     

    Results of the analysis identified significantly faster overall speaking rate and articulation rate for the NZE group compared to the AE group.