Norwegian Speed: Fact or Factoid?

« previous post | next post »

According to Heejin Lee and Jonathan Liebenau, "Time and the internet", in Hassan and Thomas, Eds., The New Media Theory Reader, "speed is contagious", and so everything is faster today, from plays and musical performances to Norwegian politicians:

The claim about musical performance seems questionable to me, but as a phonetician, I'm more interested in those Norwegian orators and their precisely measured speaking rates. However, Lee & Liebenau's paper lacks footnotes, and I was not able to find anything promising in the bibliography, and I got nothing from Google Scholar searches combining "phonemes per minute" with various keywords such Norway, Parliament, debate, etc.

Can anyone do better? Is there really a political scientist who measured the speaking rate of Norwegian politicians between 1945 and 1995? Where is this documented? Did (s)he really find an increase in average rates from 584 in 1945 to 863 in 1995? What were the other details?


  1. ambrosen said,

    September 13, 2010 @ 8:18 am

    Can we assume that they worked from the transcripts, which would have had start and finish times on them? That sounds reasonable.

  2. stboyum said,

    September 13, 2010 @ 8:18 am

    I think it refers to the distinguished Norwegian political scientist, Ulf Torgersen, who published an article called "Taletempo" in the Norwegian journal "Nytt Norsk Tidsskrift" in 1999. Unfortunately, it is neither available in English nor electronically (as far as I know).

  3. Toby said,

    September 13, 2010 @ 8:19 am

    If this turns out to be true, my first instinct is that there are so many possible confounding factors that it would be wrong to attribute the change simply to a modern taste for "faster living" (if this is indeed what the authors are suggesting — it's a bit hard to tell from the snippet above). Off the top of my head, for instance:

    – as good microphones have become available, orators may have less need to talk slowly and loudly to a large chamber over background noise, and their speech may have moved towards conversational tempo

    – the background and training of your average elected Norwegian MP may have changed over time, which may be reflected in different levels of public speaking experience and training and thus different rates of delivery

    – politicians may write more (or fewer) of their own speeches these days than they used to, or have less or more time to prepare for their performances

    – the tone and content of speeches may have become more or less technical, more or less vernacular

    – the amount and the type of media attention applied to on debates in parliament will very likely have changed over time, and speakers may be consciously or unconsciously sensitive to the changing demands of their different audiences

    By the way: In the same vein, I'm a part-time conductor as well as an amateur linguist. The general consensus among fellow conductors (based on no more than an accumulation of anecdotal evidence, I'm sure) is that music is often these days taken a little quicker than the composers intended it. (I'm always amazed to see the recommended tempos of congregational hymns in 100-year-old hymn books: they are about half the speed at which choirs sing them today.) But again, there are so many confounding factors that I'd be very hesitant about drawing conclusions. For instance, nearly all composers/editors specify their intended speed with an adjective (andante, largo, whatever) and many more recent ones specify an exact tempo (100 beats per minute). The balance between trying to divine the composer's perceived intentions and doing what "feels right" is different with different conductors; some of us do a lot of historical research before choosing the "right" tempo, others just go with the feel of the thing. Also, chosen tempo is not just an artistic choice, it's also a practical one based on the size and capabilities of the orchestra, the patience of the audience, and — very significantly — the acoustic properties of the performance venue.

  4. George said,

    September 13, 2010 @ 8:22 am

    In old newsreels, often people seem to be speaking faster than today and with herky-jerky motions. I have assumed that this is a function of the recording devices.

    I wonder if Lee & Liebenau used the same devices to compare the speech and music.

  5. stboyum said,

    September 13, 2010 @ 8:51 am

    Some information here:

  6. bulbul said,

    September 13, 2010 @ 9:50 am

    Thanks stboyum. Based on your information, I think the paper is:
    Ulf Torgersen, "Taletempo." Nytt norsk tidsskrift 1999: 16, 3-5.

  7. John said,

    September 13, 2010 @ 9:51 am

    Baseball games would seem to provide a counter-example.

  8. bulbul said,

    September 13, 2010 @ 10:01 am

    Oh crud, stupid proxy or something, I didn't see stboyum's original post. Nevermind.

  9. Sigve said,

    September 13, 2010 @ 10:12 am

    The article is also printed in Styre og stell: artikler i utvalg 1966–2002 [Government: Selected articles 1966–2002] (2005) Torgersen, Ulf & Bredo Berntsen. Oslo: Unipub. ISBN: 82-997143-0-3

  10. Michael said,

    September 13, 2010 @ 10:14 am

    As far as music is concerned, I know that there are many early 20th C recordings of popular symphonies (Beethoven, Mozart, etc.) that are much slower than mid-century performances recorded by Toscanini and von Karajan. Current tempi are the same as those. The old recordings aren't faster because of mechanical issues in playback or recording…the pitch is the same. So we are technically still playing some music slower than 100 years ago.

  11. andy said,

    September 13, 2010 @ 10:15 am

    From the link from stboyum, it seems that the researcher worked out the rate of speech from the rate of punches done by the official parliament stenographers (which transcribe in real time).

  12. Michael said,

    September 13, 2010 @ 10:15 am

    Sorry…I meant to say that those early 20th-century recordings are FASTER, not slower.

  13. a George said,

    September 13, 2010 @ 10:23 am

    @ George
    In an attempt to be viral, based on a social website that was due to close, I posted a small research note/work in progress on entitled "Two Exercises in Source Criticism – Sarah Bernhardt and Sadayakko". In this I discussed recordings from about 1900 and the way they are used today — and then. One of the most important influences is the speed with which the sound carrier moves relative to the reading head. The speed influences both tempo, pitch, and timbre. In acoustic disc, cylinder, or graphic recording the equipment (horn, conduits, soundbox, diaphragm, and levers) influence the timbre. Modern reproduction does not materially influence timbre, unless it is on purpose, but observations contemporary with acoustic recordings were obtained by means of acoustic reproduction, which does affect timbre.

    In order to keep track of these influences we apply Source Criticism, which is a discipline I introduced for sound recordings in 1981 and later developed further.

    Particularly in the case of Sarah Bernhardt it turned out that contemporary use of her recordings apparently placed the speed of reproduction at what all other approaches to the problem would consider ridiculous — much too fast. That is why I consider the note a "work in progress" – my preliminary measurements of the timings of declamations could not be the whole story.

    If my seeding of the viral distribution turns out not to work, and if another occasion arises on this website, I shall provide a more detailed text (list owners permitting)

  14. rootlesscosmo said,

    September 13, 2010 @ 11:42 am

    Dorottya Fabian, Bach Performance Practice, 1945-75 lists timings for recordings of the Brandenburg Concertos Nos. 4 and 5, as well as of the Goldberg Variations. Some relevant pages aren't available through Google Books, but her findings contradict my vague impression that tempi in Baroque performance, particularly in fast movements, had been getting steadily faster since the 1950's.

  15. Bill Walderman said,

    September 13, 2010 @ 1:00 pm

    A 1987 NYT article about Beethoven's notoriously brisk metronome markings and the tempos at which his symphonies are taken, which renders the assertion about an acceleration of tempos questionable.

  16. dw said,

    September 13, 2010 @ 1:18 pm

    There is a great website plotting a graph of the tempi of 419 recorded performances of the first movement of Beethoven's Eroica symphony here. The date range is from 1924 to 2000. The average tempo seems to have slowed down during the first half of the twentieth century and speeded up during the second half, but there's huge variation at all times.

    The claim that performances of the Pastoral Symphony today "are much faster today than the performances of 200 years ago", is ludicrous. Note that the book from which this study is taken was published in 2006, while the Pastoral symphony was not composed until 2008, so there were no performances "200 years ago" when this study was published. In addition, there is no consensus on the tempo of today's performances. For what it's worth, though, the metronome marks supplied by Beethoven are usually faster than achieved in the vast majority of performances today or throughout recorded histoey

  17. micah said,

    September 13, 2010 @ 1:47 pm

    Here's a scatter-plot of tempo vs. time for a bunch of recordings of the first movement of Beethoven's 3rd symphony. I'm not seeing any overall trend, and neither did the guy who made the plot in the first place. He does pick out some trends in various "versions" of the movement, but it's not anything nearly so simple as "tempi increase with time".

    Of course, this only goes back to 1924. I suppose it's possible that people before that liked their Beethoven slower…

  18. stripey_cat said,

    September 13, 2010 @ 2:18 pm

    My money is probably on the changing venues and media hypothesis. If you're speaking (or trained to speak) in large, echoing halls or outdoors to large rallies, you have to slow down (especially if you're un-miked). If you're being interviewed on radio or telly, you talk at a normal conversational speed. Decent amplification in a fairly soft room is somewhere in the middle. I'd suggest that in the last 60 years, the first category has become less common, and the second more common, so the medium speed (ie the one they're studying) has shifted faster.

  19. richard howland-bolton said,

    September 13, 2010 @ 2:18 pm

    dw: 1808, surely—Ludwig (or *Ludvig as SOME people would no doubt have it) was I believe 238 in 2008 and far too feeble to compose.

  20. j-g-faustus said,

    September 13, 2010 @ 2:21 pm

    Google translate from a 1999 newspaper article reporting the result, with some additional detail:

    Perhaps of particular interest is that two of the politicians agree that they today speak "too fast", and associate it with "lack of training in public speech" – i.e. a bad habit.

    By that metric perhaps the politicians of the 50's were more professional?

  21. Sigve said,

    September 13, 2010 @ 3:17 pm

    About the archives of the Norwegian parliament: there are records by stenographers of the debates from 1857, and sound recordings from 1989/1990. I haven't seen the stenographed manuscripts, so I don't know how detailed they are with respect to time indications. This article in the newspaper VG says that numbers Torgersen studied were the number of punches by the stenographers in the records, divided by the number of minutes the debates lasted.

    Here is an interesting article where a stenographer looks back on her 50 years working in the parliament.

    Det er visstnok dokumentert at dagens stortingsrepresentanter snakker 50 pst. fortere enn deres forgjengere gjorde for 50–60 år siden. – Enkelte representanter snakker nok bortimot dobbelt så fort. Dette er selvfølgelig uttrykk for en generell trend i samfunnet, men kan også ha noe å gjøre med reguleringen av debattene. Kings Bay-debatten i 1963 varte i fire samfulle dager, samtlige frem mot midnatt. Taletiden var uhyre romslig. Hovedtalere pluss en rekke statsråder hadde én times taletid, med 20 minutters replikkordskifte, og deretter var det en mengde halvtimesinnlegg. Vanlig taletid for eksempel under trontaleordskifter var 45 minutter. Man kunne med andre ord ta seg tid. I våre dager med stram taletid og mye man anser som viktig å få sagt, er det vel nødvendig å speede opp for å få servert alt som ligger en på hjertet.

    It has apparently been documented that the speech rate in the parliament has increased 50% in the last 50–60 years. Some representatives speak twice as fast. This is of course a reflection of a general trend, but it could also be related to the form of the debates. The Kings Bay debate in 1963 lasted four days, each going on until midnight. The representatives were generously alloted speech time. The main speakers and members of the governnment could speak for an hour, while responses could be up to 20 minutes. Then other speakers followed, who could speak for 30 minutes. Normal speech time during the opening debates was 45 minutes [the debates Torgersen examined]. Today the speech times are much shorter, and when there is a lot to be said about an issue, I guess it’s necessary to speed up the speeches to deliver everything that’s on one's mind.

    Another comment that is interesting, but perhaps not relevant for Torgersen’s study, is the following:

    De to talerstolene som nå er plassert midt på gulvet til bruk for spørretimer og replikkordskifter, er selvfølgelig en rasjonell nyskapning. Det sparer en masse gangtid, og spørsmålene og replikkene kommer i mitraljøsetempo; det er nærmest som man snakker sammen over bordet eller i et TV-studio. Forsvunnet er den tenksomme, ofte overdrevet langsomme vandringen over gulvet mens man tydelig kunne se at det ble foreberedt en replikk som virkelig skulle sitte. «Nerven» er ikke der på samme måte som før.

    The two podiums placed in the middle of the room for questions and responses are a rational innovation. It saves a lot of walking time, and the questions and comments come in machine gun tempo; it's almost as if they are talking together across a table or in a TV studio. The thoughful, often exaggaratedly slow walk across the floor while preparing for a comment is gone. You no longer have the same “nerve”.

    If Torgersen had studied regular parliamentary sessions, I would have thought that it might be necessary to control for speaker age. The youngest member of the current parliament was 23 years old in her first year, something I strongly doubt would have been the case in 1945. But since Torgersen seems only to have studied the opening debates, where the party leaders probably do most of the speeches, speaker age probably wouldn't be that variable.

    Anyway, it would be cool if Lee & Liebenau have this directly from the article in Nytt norsk tidsskrift [The new Norwegian journal]. But maybe they have it from some other source in English citing Torgersen.

  22. j-g-faustus said,

    September 13, 2010 @ 4:51 pm

    Quick comment on the Google translation of the above page:
    "Even select Payday speak at a lower pace"

    should be
    "Lønning himself chooses to speak more slowly" (Lønning is a name, but can also mean "paycheck" [don't know where Payday came from], similar to how Smith can mean "metal worker")

    For a possibly similar phenomenon in English: Doesn't Martin Luther King in "I have a dream" speak more slowly than what is common on TV or radio today?

    Like earlier posters, I would bet it is due to the combination of
    a) change of venue (outdoors and large halls less common today)
    b) improved technology
    c) (lack of) training in public speaking

    Assuming the numbers are correct (I haven't seen the study, but the methodology described in the newspaper article seems reasonable), I'm guessing that the low rate is similar to how theater actors speak, while the high rate is closer to normal face-to-face conversation.

    In short, I think the reported speed difference sounds plausible, but I can't immediately see that it can be extrapolated to "the conversational speaking rate has increased" or "everything is faster today".

  23. KevinM said,

    September 13, 2010 @ 5:23 pm

    richard howland-bolton wrote "dw: 1808, surely—Ludwig (or *Ludvig as SOME people would no doubt have it) was I believe 238 in 2008 and far too feeble to compose."

    Indeed – by then, he was fully decomposed.

  24. Dw said,

    September 13, 2010 @ 5:37 pm

    @richard howland-bolton, @KevinM:

    Not for the first time, I curse the lack of an "Edit" or "Delete" button for LL comments… :)

  25. Liam said,

    September 13, 2010 @ 6:01 pm

    The whole sentence:

    "I have no way of knowing whether performances of Beethoven's sixth symphony, the idyllic Pastoral symphony, are much faster today than the performances of 200 years ago."

    I don't know what precedes this sentence (some skillful googling could perhaps uncover it, one sentence at a time), but this one alone doesn't really make any claims.

  26. Liam said,

    September 13, 2010 @ 6:02 pm

    Or, if anyone has the actual book, I'd be curious to see what else is on page 272.

  27. Dw said,

    September 13, 2010 @ 6:24 pm


    I'm in awe of your Google Books skills. How did you find the beginning of that sentence?

  28. Liam said,

    September 13, 2010 @ 7:07 pm


    The preceding page (272) wasn't included in the preview, but, by searching for words that seemed likely to be present, I was able to get the rest of that one sentence. Beethoven's or symphony may have been what did it.

  29. Dw said,

    September 13, 2010 @ 7:18 pm


    Ahh — very clever!

  30. Peter said,

    September 13, 2010 @ 8:04 pm

    Following Liam's technique gets us a few sentences further back:

    Ramonet claims that during the last 30 years, more information has been produced than during the previous 5000 years! He illustrates the point with an example: 'a single copy of the Sunday edition of the New York Times contains more information than a cultivated person in the eighteenth century would consume during a lifetime'. I have no way of knowing…"

    but I can't chase it any further!

  31. Jason L. said,

    September 13, 2010 @ 11:59 pm

    In Beethoven's time, fixed-pitch instruments (keyboard instruments, fretted instruments, sorta-kinda woodwind instruments) were tuned such that commonly-used keys were more in tune and uncommonly-used keys were less in tune. Stringed instruments and human voices of course lack inbuilt punctuation in their pitch palette, but the prevailing technique of the time (extended sixth-comma meantone) resulted in harmony that was more restful and less unstable and active than what has been the norm since the late Romantic (equal temperament).

    I once heard it remarked that Western music is fast because it is out of tune. Major and minor triads today (the harmonic foundation of Western music — folk, pop, "art", you name it) feel less stable than they did in Beethoven's time. So perhaps changes in tuning, which gradually resulted in less stable, more active harmony from Beethoven's time to the turn of the century but have remained largely constant since then, contribute something to the preference for faster tempos.

    A somewhat related phenomenon may be that pitch has slowly gone sharp over the centuries. In the Baroque era, the A above middle C was about 415 Hz; today it's 440 or 442 Hz. Beat rates in intervals that deviate from "just intonation", that is, small integer ratios between notes, like 5:4 for a major third, are faster the higher you go. An interval that's off of a small integer ratio by, say, 5 Hz, near middle C, will be produce beats of 20 Hz two octaves higher. So deviations from just intonation that are tolerable at lower pitches can become intolerable at higher pitches. Thus, mapping the tonic tone of Beethoven's 7th Symphony (in A major) to 440 Hz rather than something like 420 Hz will result in faster beat rates for out-of-tune intervals, making the same music sound more out of tune today than two hundred years ago.

  32. a George said,

    September 14, 2010 @ 5:22 am

    @ Jason L.:
    Ross W. Duffin published a very thought-provoking and highly recommendable book in 2007: "How Equal Temperament Ruined Harmony (and Why You Should Care)". He mentions two CDs "Beethoven in the Temperaments" and "Six Degrees of Tonality", performed on grand pianos tuned according to various temperaments and performed by Enid Katahn. These are practical experiments on the listener.
    While it is true that the standard pitch has risen generally and gradually in the capitals of the countries where Western art music was played, there were deviations that could perhaps serve as experiments seen from today. The Philharmonic Pitch in England (later demoted to Old Philharmonic Pitch) had a standard a at 452 Hz (and higher) from 1846 until the New Philharmonic Pitch was introduced in 1896. However, British military bands continued the use of the Old Philharmonic Pitch until 1927. You will hence find records of Nellie Melba where she sings to Old Philharmonic Pitch accompanied by H.M. Coldstream Guards in 1905 and plenty of others where she sings to the new pitch.

    Another recommended book is Mendel, A. (ed): Musical Pitch. Monographs by Alexander J. Ellis and Arthur Mendel, Frits Knuf, Amsterdam 1968. Not the least because it reprints Guido Adler's review of Ellis from 1888.
    Those concerned with correct reproduction of historical records really have puzzles to solve; at about the time the military bands tuned down, the Columbia Graphophone Co. decreased the speed of their records from 80 rpm to 78 rpm. That fraction is the same as the fraction between the tuning pitches! It could mean that the old Columbia recordings were still "valid", albeit faked.

  33. maidhc said,

    September 14, 2010 @ 5:25 am

    In the Renaissance, pitches were not standardized, but were generally at A=440 up to maybe A=450.

    In the Baroque era, pitches were still not standardized, but tended to go somewhat lower.

    In the 20th century, revivalists decided that Baroque pitch was A=415. This is useful because it is a semitone lower than the modern standard A=440. Not that this was any kind of standard in the Baroque era.

    Baroque pitch acquired a superior social standing among 20th century performers, such that when playing Renaissance music, people would say "I normally play Baroque music, but when I'm slumming I play Renaissance music, but don't expect me to retune to anything other than A=415."

    I live in the San Francisco Bay Area, where we have the San Francisco Early Music Society, which puts on concerts and so forth. Almost all of their presentations are Baroque music. Medieval and Renaissance music are almost invisible.

  34. richard howland-bolton said,

    September 14, 2010 @ 7:13 am

    maidhc@ 'Medieval and Renaissance music are almost invisible.'

    Yup those guys put the 'late' in early music, and this is sad since (as we all know) the high point of Western music occurs half of way through the Credo of Josquin's Missa L’homme Armé (a sexti toni).
    Here's a scholarly essay that explains it all:

  35. Mike Maxwell said,

    September 14, 2010 @ 11:33 am

    richard howland-bolton wrote "dw: 1808, surely—Ludwig… was I believe 238 in 2008 and far too feeble to compose." And KevinM commented that "by then, he was fully decomposed." On either account, I'm sure he was also deaf by then, which surely would have made composing music impossible, right?

  36. richard howland-bolton said,

    September 14, 2010 @ 12:13 pm

    Deafness didn't seem to stop him before he died!

  37. peterv said,

    September 14, 2010 @ 7:13 pm

    Regarding the act of discerning a dead composer's intentions from the written score: Until the middle of the 19th century, solo pianists played most left-hand chords as arpeggios (ie, asychronously). Composers of solo piano music, however, notated these chords almost always as synchronous chords, with all notes to be played simultaneously. Performers apparently understood, and composers apparently knew that performers understood, that the written instructions were not be followed strictly. This difference between notation and performance style was true even for those romantic composers of solo piano music (eg, Beethoven, Mendelssohn, Chopin, Liszt) who were themselves also well-known performers of solo piano music.

    So the tempo indications that a composer such as Beethoven added to his notated scores may or may not have borne any close relationship to the tempi at which the music was actually played, even when the composer him- or herself was performing or conducting. Mendelssohn, for example, according to Wagner, played everything very fast, regardless of the tempo indications on the score.

  38. Chris Vosburg said,

    September 15, 2010 @ 2:10 pm

    I've always assumed that Rossini, the heavy-metal headbanger of classical composing, had at least one proprietary tempo marking: "Let's see how fast you can make the orchestra go, Signor Conductor".

  39. pm said,

    September 16, 2010 @ 8:24 am

    The current Australian Prime Minister – notorious for her deliberative speaking style – is presumably an exception to this alleged trend towards ever-faster pollie-speak.

RSS feed for comments on this post