“What Are The Wild Waves Saying?”: Logograph Records of Spoken English, 1873-1878

Before Thomas Edison unveiled his speaking phonograph in 1877, another instrument had already begun to make automatic records of spoken language.  The concept was quite simple: someone would talk into a funnel with a membrane at the other end while a stylus attached to the membrane traced its motions onto a moving sheet of paper as a wavy line.   The records it made weren’t intended for playback; instead, people were expected to learn to read them visually, either as a means of studying speech in a new way or as a substitute for manual shorthand that would soon put stenographers out of work.  In the end, nobody succeeded in deciphering actual words from the records, so the stenographers got to keep their jobs for a while longer; but when the Bell telephone and Edison phonograph came along soon afterwards, commentators pointed to this earlier invention as their “forerunner.”[1]

If you’re familiar with the early history of sound recording, you might assume that I’m writing about the phonautograph of Édouard-Léon Scott de Martinville, which inscribed airborne sound waves onto soot-blackened sheets of paper as wavy lines analogous to the grooves of LPs.  But in fact the invention I have in mind here is the logograph of William Henry Barlow (1812-1902).  Like Scott, Barlow wanted to convert speech into a legible form of writing; but unlike Scott, he sought to accomplish this by recording breath impulses rather than sound waves.  Hold your hand up close to your mouth and speak: “Peter Piper picked a peck of pickled pepper.”  Scott’s goal was to record the acoustic phenomenon you hear; Barlow’s goal was to record the pneumatic phenomenon you feel.  That’s an important difference.  But there’s another important difference besides. Scott was Parisian and made all but one of his known spoken-word recordings in French (the exception is in Italian), while Barlow was British and made most of his recordings in English.  So Barlow’s recordings may not count as “sound recordings”—I’ll have more to say about this later—but they are the oldest known automatically-inscribed records of speech in the English language, dating back in some cases to 1873 or 1874.  Even if we can’t reproduce the sound of that speech as such, we can still listen to the records it left behind to experience some aspects of that speech in a meaningful way.

Barlow was a prominent British civil engineer known mainly for his work on railroad infrastructure.  The logograph was very much a sideline for him, perhaps even a hobby of sorts, although he took it seriously enough to publish and present on it in some fairly prominent venues, and the reputation he had acquired in civil engineering may have led his contemporaries to pay more attention to his efforts to record speech than they otherwise would have.  He described his original ambition and inspiration for the logograph during a talk he gave in 1878:

This instrument…owes its origin to an attempt to make something which would do short handwriting instead of employing the services of the man who sits there [he must have pointed to the stenographer who was busy taking down his words]….  Everybody knows that, in speaking, the air contained in the lungs is sent forth from the mouth in spurts of different magnitude and with different degrees of velocity and intensity.  I may say that forty years ago I first thought of this when I saw an old Turk smoking his pipe.  The Turks, as you know, inhale the smoke into their lungs, and, as they speak, you can see all the action of the air coming from the mouth.  I was also informed of the extraordinary sensitiveness of some of the instruments used for registering indicator diagrams of the steam engine at very high velocities, and it occurred to me to put these two things together and form a delicate apparatus, and so obtain indicator diagrams showing the relative powers of the different syllables uttered by the human voice.[2]

Barlow’s anecdote about the “old Turk” must have dated from his employment in Constantinople during the mid-1830s, when he’d worked on erecting “works and machinery for re-casting and re-boring Turkish Ordnance,” as we learn from his obituary.[3]  The registering indicator diagram was a mechanically produced graph of pressure against volume that was used to gauge the work done by steam engines; Barlow would have encountered this industrial recording technology through his professional involvement with steam locomotion.  So those diagrams had shown him that it was possible to record rapid and minute fluctuations in the volume and pressure of steam, while the puffs of smoke coming from the mouth of the “old Turk” had illustrated vividly to him that spoken syllables were accompanied by distinctive fluctuations in the volume and pressure of air.  With those two conceptual models in mind, he’d set about devising a means of recording the pneumatic actions associated with speech.  Note that there’s nothing here about any background knowledge of acoustics or the physiology either of hearing or speaking—just smokers and steam engines.

Scientific Proceedings of the Royal Dublin Society 2 (May 1879),153.

Barlow reportedly made “some very interesting experiments” in pneumatic speech recording during 1873,[4] and he gave his first public account of them in a paper received by the Royal Society on February 23, 1874—and officially read on April 16—entitled “On the Pneumatic Action which accompanies the Articulation of Sounds by the Human Voice, as exhibited by a Recording Instrument.”   The device he had come up with consisted, first, of a speaking trumpet leading to a tube that widened out to a 2¼-inch aperture covered with a membrane of goldbeater’s skin or gutta percha.  The tube was provided with a hole to allow the air coming in through the speaking trumpet to escape rather than building up inside it.   On the other side of the membrane was a spring (or, in later accounts, a stud) which pressed lightly against it “to prevent as far as practicable the effects of jar and vibration.”  Attached to that spring was a lightweight aluminum arm that held a “marker”: a tiny brush protruding from a glass tube and fed with color.  Finally, a strip of paper was “made to pass under the marker in the same manner as that employed in telegraphy.”[5]  Barlow had been in charge of setting up telegraph lines in connection with his railroad work, so he knew all about such things.[6]

The Practical Applications of Electricity (London: Institution of Civil Engineers, 1884), 29.

So that was the gadget which Barlow brought to the attention of the Royal Society as its “inventor and exhibitor”[7] in the spring of 1874.  He didn’t give it any particular name in his paper—it was just a “recording instrument”—but when the Popular Science Review reprinted the same paper to bring it before a larger audience, it bore a new, catchier title: “The Logograph, or Writing by the Voice.”[8]  A “logograph” is literally a “word-writer,” and that was apparently what Barlow thought he had invented, albeit with certain qualifications:

[I]t will be seen that although there are instances in which considerable differences in sound do not make much variation in the diagram, yet, as a rule, every change of sound or articulation produces a change in the diagram, and that there are pneumatic actions revealed by this instrument which are imperceptible to ordinary observation.[9]

To be clear, Barlow’s logograph of 1874 wasn’t a sound-recording instrument, as Scott’s phonautograph was.  Both instruments recorded the motions of membranes over time, but Scott had optimized his equipment for recording the vibrations of sound waves passing through the air, whereas Barlow had applied a spring to his membrane expressly to eliminate vibrations in favor of breath impulses—a combination of air volume and pressure.  Nevertheless, the logograph did make records that documented actual utterances of words and sentences in the English language, and those records have the same kind of direct causal relationship between record and thing recorded as sound recordings and photographs do: the speech actually caused the records.  These records don’t contain enough information, or the right kind of information, for us to decipher words visually or reproduce them intelligibly.  However, they did capture some nuances of articulation with photographic accuracy, as Barlow later pointed out:

The CHAIRMAN [Latimer Clark] asked whether, in repeating the same words over and over again, the signs recorded are constant.

Mr. BARLOW: They are constant, subject to very minute variations that arise in this kind of thing.  It is almost impossible to speak the same words exactly in the same manner each time.  The instrument is purely phonetic, and however minutely your articulation differs at one time from another that difference is registered.[10]

Those minute differences corresponded to actual utterances of English words during the 1870s, no two ever alike; and, as far as we know, this was the first time anybody had ever tried to make automatic records of someone speaking in the English language.  The rumor that Scott recorded Abraham Lincoln’s voice on a phonautograph in the White House during the 1860s is wholly unfounded.[11]  Franciscus Cornelis Donders used a phonautograph in Utrecht to record the British phonetician Henry Sweet uttering English-language vowel sounds on February 27, 1874,[12] but that was four days after Barlow had submitted his paper to the Royal Society, and in any case, these weren’t recordings of complete words, much less phrases or sentences.  Charles Anson Morey, a student at the Massachusetts Institute of Technology, similarly published a phonautogram of the vowel “oo” in “mood” in August 1874, probably recorded sometime earlier that year.[13]  That’s it.  Nothing else purporting to be a record of English-language speech is known to exist—or even to have been made—in 1874 or before.  And Barlow didn’t confine himself to tracing individual speech sounds; he recorded whole phrases and even pieces of poetry.

Even though we’re dealing with records of breath impulses rather than sounds, we can still play them back in the sense of educing them—making their content accessible to our senses from a state of latency—as phenomena unfolding over time.  One way to do this would be to use the records to control the amount of air escaping through a jet so that we could feel it blowing against our hands, “reproducing” the original breath impulses, but that would require some rather specialized playback equipment.  What I’ve done here has been, instead, to sonify Barlow’s records using white noise in a way that I believe gives the impression of air escaping through a jet.  First, I’ve used the value of the y axis, corresponding to the force which the air originally exerted on the logograph membrane, to control the overall volume of noise.  Where the curve is higher, corresponding to less force, there’s more noise; and where the curve is lower, corresponding to less force, there’s less noise.  Second, I’ve tied the increase in volume to an increase in frequency range, such that as the curve goes up, corresponding to more force, the noise gets both louder and higher in pitch.[14]  In this way, I’ve actualized Barlow’s breath-impulse records as audible fluctuations in hiss.  I’ve also prefaced each example with an announcement of whatever word or words it’s supposed to document, taking language-learning tapes as my model.  So you’ll hear me say some words, and then you’ll hear Barlow’s logograph record of those same words, and so on and so forth.

The introductory paper Barlow submitted to the Royal Society was accompanied in print with illustrations of various records he’d made on his logograph sometime during 1873 or early 1874, and I’ve educed all the examples that represent actual English words, phrases, or sentences (which is most of them).  He used his first few examples to show the similarities in records made when the same word was spoken twice by the same person:

Experimental Tests.
Experimental Tests.
Recording Instrument.
Recording Instrument.
Repetition Trials.
Repetition Trials.

He next drew attention to what he called the “silent discharge,” meaning a discharge of air that the logograph recorded after a syllable that ended with certain consonants, especially plosives.  Put your hand close in front of your mouth and say the word “let”; the second puff of air you’ll feel is what Barlow had in mind here.  In fact, the second puff of air is no more or less silent than the first, but Barlow was thinking in terms of recording syllables, and in his analysis he assimilated the first puff of air to the syllable that followed it.   By contrast, the “silent discharge” seemed to him to create an extra bump in his records—a bump that didn’t correspond to any audible syllable.  A single-syllable word thus ended up with two bumps, a two-syllable word with three bumps, and a three-syllable word with four bumps:

Let.  Attempt.  Integrate.

However, Barlow observed that this silent discharge, which he marked with a d in his diagrams, was “under the control of the will; for by holding the breath immediately after pronouncing the word, this part of the diagram can be altered and the discharge of air postponed or let off gradually”:

Let.   Attempt.  Integrate.

And he noted further that if “another syllable be added to each word, making them terminate with consonants of softer sound, the air which would have been silently discharged is used to form the syllable added, and the subsequent silent discharge is very much diminished”:

Letter.  Attempting.  Integrating.

Barlow detected some additional “silent or, rather, insensible actions” within words, as he illustrated with the word excommunicate: “it is seen that the part p, which is the secondary sound of the syllable ‘Ex,’ becomes compressed, its length being shortened and its height increased; so that although nearly insensible as regards sound, it becomes developed into the form p’, and constitutes the most prominent feature of the diagram when the whole word is pronounced”:

Ex.  Ex.
Communicate.  Communicate.
Excommunicate.  Excommunicate.

Sometimes, Barlow observed, a word was measurably shortened when another syllable was added to it:

Strength.  Strength.
Strengthen.  Strengthen.

To conduct a “test of the rapidity of action of the instrument,” he had recited a familiar tongue-twister “at the rate of six syllables per second; and it will be observed that there are two principal upward and two principal downward movements to many of the syllables, besides other subsidiary actions.”  This example also gave me the speed reference I’ve used in my eductions.[15]

Peter Piper picked a peck of pickled pepper.
Peter Piper picked a peck of pickled pepper.

Barlow distinguished “words of quantity,” which appeared to entail the discharge of an unusually high quantity of air—

Voice.  Force.  This.  Hear.
Voracious.  Ferocious.  Sufficeth.

—from “words of intensity,” which had a higher “degree of compression”:

Tick.  Peg.  Kite.  Tin.
Attended.  Tentative.  Deducting.

“The difference in the action between whispered sounds and those spoken loud,” he wrote, “is not so great as might have been expected”:

Incomprehensibility. [whispered faintly]
[whispered forcibly]
[ordinary tone of voice]
[spoken loudly]

Finally, Barlow presented two different records of a verse from Thomas Campbell’s poem “Hohenlinden,” the first recited continuously, and the second uttered as a sequence of isolated words:

By torch and trumpet fast arrayed,
Each Horseman drew his battle blade;
And furious every charger neighed,
To join the dreadful revelry.

his—battle—blade; —And

“It will be observed,” he wrote, “that the diagrams of the separate words, although they become modified when grouped together, are more or less discernible in the lines continuously spoken; and the similarity of sound at the termination of the first three lines, which constitutes the rhyme of the verse, is represented in the similarity of form, or in the character of the form, of the terminations of the diagrams of these three lines.”  This was as far as Barlow had taken the idea as of the spring of 1874.  He offered his preliminary results at that time merely as a proof of concept, “to show that the articulation of the human voice is accompanied by definite pneumatic actions, and that those actions, many of which are insensible to ordinary observation, are capable of being recorded.”[16]

Barlow’s paper received moderate attention in reports of new scientific developments.  The loan collection of scientific apparatus at the South Kensington Museum had acquired a logograph by 1876,[17] and about that same time Francis Galton experimented with one enough to observe in a letter to Charles Darwin that “the traces were greatly modified under different conditions of cadence.”[18]  Barlow himself doesn’t seem to have published or presented anything about the logograph for a few years after 1874, but on February 27, 1878, he gave a talk about it for the Society of Telegraph Engineers, complete with a demonstration of recording;[19] and on March 18, a new paper he’d written about it was read on his behalf before the Royal Dublin Society by physicist and parapsychologist William Fletcher Barrett.[20]  The Bell telephone and Edison phonograph had then recently sparked a general fascination with all aspects of the transmission and recording of sound, which might account in part for the new surge of interest in the logograph.  But Barlow had also experienced a technical breakthrough of his own, as he explained:

Having made these experiments, which resulted in showing the peculiarities of articulation, and how the different letters were formed, and what happened to them, it occurred that the marker I was using became fixed to the disc.  That arose from having used gold-beaters’ skin, which caused the attachment of one to the other.  I found the consequence was to produce a vibration from the vowel sounds.[21]

Sometime between 1874 and 1878, then, Barlow had discovered—apparently by accident—that his logograph could be made to record vowels as well as consonants if he fastened either the stud or the marker itself to its membrane.  However, doing this also reduced the overall sensitivity of the instrument, which in its original form was supposedly so great “that in a long sustained action of breath the pulsations of the heart are shown.”[22]  For that reason, Barlow attached the marker to the membrane only when making studies of vowels, and even then he acknowledged that the logograph could show only whether vowels were present, since “the scale of the diagram is too small, and the arrangement of the instrument is not adapted, to show those changes which distinguish one vowel sound from another.”[23]  I’ll return to the recording of vowels and its implications shortly.

The new paper Barlow had written for the Royal Dublin Society was illustrated with many previously unpublished logograph recordings—so many, in fact, that I’ve been somewhat selective about the ones I’ve included here.  Some of them he used as evidence in support of arguments about the nature of speech articulation, and one subject he addressed in this way was the “jointing” of syllables by consonants.  Sometimes, he wrote, the consonant at the beginning of a syllable was similar enough to the one at the end of the preceding syllable to allow for a “simple re-action”:

Trigonometrical.  Incompatibility.

In other cases, by contrast, “the position of the organs of the mouth at the end of one syllable is not suited to commence the next, and a change of action has to take place which requires more time”:


Obtain.  Subtend.
Sackbut.  Captain.

Barlow claimed that such words effectively contained a “suppressed syllable”—ob-bit-tain, sub-bet-tend, sack-keb-but, cap-pet-tainwhich was sometimes “bridged over” by the letter s:

Extinct.  Abstain.

He also detected a “tendency to re-acting upon one consonant,” leading to “compounds which we are not accustomed to consider as such”:  uninterrupted-dly, night-tly, gallant-tly.

Uninterruptedly.  Nightly.

Another subject which Barlow tried to address in his Royal Dublin Society paper was variations in accentuation, which seemed to have vaguely predictable effects on the spacing of logographic bumps:

Attitude.  Attribute.  Majesty.
Attention.  Attribute.  Majestic.



But together with all these experimental records of individual English words, which were interspersed in the main text of the article, Barlow also included a few records of connected phrases on a plate at the end.  The first of these was an excerpt from Thomas Campbell’s poem, “Ye Mariners of England: A Naval Ode”:

Britania needs no bulwarks, No towers along the steep, Her march is on the mountain wave, Her home is on the deep.

Next came “What Are the Wild Waves Saying?”—the title and first line of a well-known song with music by Stephen Glover and lyrics by Joseph Edwards Carpenter that had been based on a passage in Charles Dickens’s Dombey and Son:

What are the wild waves saying.  What are the wild waves saying.

Then, whispered:

The Lilies of the valley.  The Lilies of the valley.

And finally:

The vibratory action is now very decided.
The vibratory action is now very decided.

The published version of Barlow’s talk for the Society of Telegraph Engineers contained a similarly extensive trace of an excerpt from Alexander Pope’s “The Riddle of the World,” printed vertically along the side of one page:

Know then thy self
Presume not God to scan
the proper study of mankind is man

This last record may have been the example Barlow had pointed to during his talk when discussing the logograph’s documentation of vowel sounds:

These larger lines represent the consonant actions.  You observe that following them is a fine vibration, and that vibration is the vowel sound, or rather I believe there is a certain organ which vibrates whenever the vowel sound is uttered.  I think it is a kind of drone which is present whenever the vowel sound is uttered.[24]

Whether or not the record of the Pope couplet was the specific one Barlow indicated as he was saying this, it clearly displays the features he was describing—namely, “large” curves corresponding to consonants combined with “fine” ones corresponding to vowels.  The same is true of the “what are the wild waves saying” and “the vibratory action is now very decided” records found in the Royal Dublin Society paper.  Barlow had apparently carried out some experiments with singing to figure out how best to make records of this kind:

If the gamut be sung into the mouth-piece, although more than one note will produce vibration, yet its greatest development is found at some one note, which may be assumed that corresponding with the vibration rate of the membrane in its then state of tension; and the best exhibition of vowel action is obtained by speaking in a monotone upon such note when discovered by trial.[25]

So in all three cases—“know then thyself,” “what are the wild waves saying,” and “the vibratory action is now very decided”—Barlow had probably identified a resonant frequency of the logograph membrane and then spoken the phrase to be recorded in a monotone on that note (or had someone else speak it that way).  This wouldn’t have sounded much like natural speech.

The existence of records such as these forces us to reconsider the relationship of Barlow’s logograph to the history of recorded sound.  Back in 1874, the logograph had plainly recorded breath impulses, and not sound waves.  Even in 1878, it mostly recorded breath impulses, and that’s the way posterity has generally remembered it.  The following year, Silvanus P. Thompson confidently wrote that logograph records “do not correspond, strictly speaking, to sounds at all, and do not represent sonorous vibrations; they give the mechanical displacements of the air due to the change of wind-pressure in the cavities of the mouth during articulation.  Vowels and musical sounds made no trace at all in the Logograph.”[26]  By 1878, however, Barlow was using his logograph to record vowels and musical sounds, as we’ve seen.  And whenever it did this, it was recording sound vibrations.  That technically made it a sound-recording instrument, even if it wasn’t a very good one.

The distinction would have been lost on many of Barlow’s contemporaries, who evidently thought that the logograph had been recording speech sounds all along.  When William Henry Preece gave a lecture about the telephone before the Royal Institute of Great Britain on February 1, 1878, for example, he had this to say:

Mr. W. H. Barlow, C. E., produced before the Royal Society in 1874 his logograph, which recorded in varying lines and curves spoken language.  Here is such a line, which records the pitch, loudness, and form of the sounds emitted by the lips of the speaker, and reproducing all the elements of the voice.[27]

The minstrel boy to the war is gone,
In the ranks of death you’ll find him;
His father’s sword he has girded on,
And his wild harp slung behind him.

The “line” was a logograph record of the opening lyrics of Thomas Moore’s Irish patriotic song, “The Minstrel Boy.”  When Preece showed the same specimen in a talk about the phonograph on May 8, he again claimed that its curves gave “in their amplitude, the loudness of the sound; in their length, the pitch of the sound; and in their form, the quality of the sound—the timbre, as the French call it.”[28]  But this was strictly a breath-impulse record.  It didn’t contain any information about pitch or timbre, and it certainly didn’t reproduce “all the elements of the voice.”  As far as Preece was concerned, however, it had documented speech in essentially the same way that the telephone transmitted speech and the phonograph reproduced it.  And he wasn’t alone in this understanding.  If we read nineteenth or early twentieth-century British articles and accounts of lectures on the Bell telephone and Edison phonograph, we find Barlow’s logograph cited again and again as a direct precursor that had anticipated them as a technology for picking up sound waves out of the air.  Some commentators credited Barlow and his logograph with ideas and accomplishments that properly belonged to Édouard-Léon Scott de Martinville and his phonautograph:

Mr. Barlow argued that the voice, which sets up waves in the air, ought by these waves to be able to set a diaphragm in vibration, and that this vibration ought to be constant for all repetitions of the same syllables and words, and that if those vibrations could be caused to transmit their character to a travelling band of paper, it would be possible upon looking at the paper, having once learned the characters, to read off at any time that which had been said.[29]

Others stated that Scott’s phonautograph and Barlow’s logograph were really just two names for the same thing,[30] or even that the logograph was an improvement on the phonautograph:

In 1858 Leon Scott invented the phonautograph….  In its first form this instrument gave very imperfect tracings, but it is of great interest as being the forerunner of the phonograph….  Then came the logograph of Barlow in 1876….  This gave more accurate tracings, that is to say, tracings that indicated with more precision the variations of pressure on the membrane.[31]

In short, there seems to have been a surprising amount of confusion about the distinction between a breath-impulse recorder and a sound-vibration recorder in the first place.  Perhaps Barlow was eminent enough as a civil engineer that, in the dawning era of telephones and phonographs, his peers were eager to credit him and his invention with more prescience than they warranted.

But when we run into logograph records that contain vibratory traces of vowel sounds superimposed on the breath impulses, the boundary between sound recordings and pneumatic recordings becomes more legitimately unclear.  Preece gave another talk about telephones and phonographs at the Royal Engineers Institute on March 14, 1878, during which he showed and discussed the same logograph record of Pope’s “know then thyself” couplet that was used to illustrate the published version of Barlow’s talk for the Society of Telegraph Engineers, complete with vowel vibrations:

We have first those rapid little vibrations indicating the pitch, which are too minute to be seen.  In the prominences and valleys, the hedges and ditches, you have the amplitude of those sounds, producing their loudness, and in the particular form that these curves take you have that quality which distinguishes one word from another, or one voice from another…. Now, Mr. Edison, instead of recording these marks on paper thought he would record them on tinfoil [in his phonograph]…. and if we were to take a section of the tinfoil in the circumference of the circle it describes, we should find that the tinfoil would be indented in the same way as the curves are made on Barlow’s logograph.[32]

This time, Preece’s identification of logograph records with sound recordings is more plausible.  And if he were right, that would be significant.  Barlow’s breath-impulse traces from 1873-74 already appear to be the oldest automatically-produced records of English speech.  But if some of his later breath-impulse traces—the ones that show vowel sounds—also “count” as sound recordings, they could well be the earliest known surviving sound recordings of speech in the English language.  We don’t know exactly when they were recorded, but they were published in February and March of 1878.  For comparison, the oldest intact recordings of English-language speech otherwise known to exist are a tinfoil phonograph record made in St. Louis on June 22, 1878, including nursery rhymes,[33] and a set of photographic sound-wave records made by Eli Whitney Blake, Jr., and published during July 1878, including the phrase “how do you do.”[34]  Barlow’s records could beat those contenders by several months.

With all this in mind, I’ve educed each of the three logograph records containing vowel sounds a second time, not by simulating the hiss of escaping air with a white noise sample as before, but now as regular sound recordings, in order to test the hypothesis that Preece was right about the curves resembling the records made on Edison’s tinfoil phonograph.

What are the wild waves saying.
What are the wild waves saying.

The vibratory action is now very decided.
The vibratory action is now very decided.

Know then thy self
Presume not God to scan
the proper study of mankind is man

I’d hoped at least to hear some hint of intonation, as we do with Scott’s spoken-word phonautograms, but I was disappointed: the vowels sound more like static than drones because the wavy lines that are supposed to represent them are so irregular in form and spacing.  At first I thought that the engraver must simply have done a sloppy job with them, and that if I could track down Barlow’s original tracings I might have better luck.  But on further investigation I now doubt that’s the case.  Barlow wrote that eliminating the aluminum arm from the logograph and fastening the marker directly to its membrane gave the “greatest development” of the vowels, but that when this was done, “the lateral action which arises when the line of motion of the marker is not controlled by the arm” tended to create distortions in the record.[35]  In other words, the marker was no longer constrained by the arm to move perpendicularly relative to the direction of motion of the recording surface but could instead be tossed around in any which way by the flexing of the membrane.   So the chaotic state of the “wild waves” is likely to reflect a real problem inherent in the original recording, and not just sloppiness in how it was engraved.

Even a very poor sound recording is technically still a sound recording, but these examples plainly belie Preece’s claim that Barlow’s records documented “all the elements of the voice.”  And that should come as no surprise.  The logograph approached the recording of speech in a fundamentally different way than telephones and phonographs did.  Except for the few exceptions we’ve considered, Barlow did his best to exclude sound vibrations from his records, noting that “whispered diagrams are generally more perfect than those which are spoken” because of “the disturbing influence of the vibratory action of the vowel sounds” which he associated with voicing.[36]  On the other hand, modern sound engineers typically try to avoid popping p’s and hissing s’s, which were precisely what Barlow was most eager to capture—in fact, it might be useful to think of the logograph of an instrument designed to record only popping p’s, hissing s’s, and so forth.  It may have bequeathed us the oldest recorded traces we have of actual English-language speech, but it didn’t record that speech in anything like the way the phonograph did.  And the fact that so many writers and lecturers at the time couldn’t tell the difference suggests that an understanding of the principles behind modern sound media were slower to emerge than we might think.

This essay and its accompanying audio were originally prepared in the spring of 2014 for a publication that seems to have fallen through—at least, I haven’t heard anything about it since then.  I’m posting it here as part of an effort to free up a backlog of past work that’s been sitting around gathering digital dust.


[1] “Obituary: William Henry Barlow,” Minutes of Proceedings of the Institution of Civil Engineers 151 (1903): 388-400, at 398.

[2] “The Logograph,” Journal of the Society of Telegraph Engineers 7 (February 27, 1878): 65-8, at 65.

[3] “Obituary,” 389.

[4] “Obituary,” 398.

[5] W. H. Barlow, “On the Pneumatic Action which accompanies the Articulation of Sounds by the Human Voice, as exhibited by a Recording Instrument,” Proceedings of the Royal Society of London 22 (Apr. 16, 1874): 277-86, at 277-278.

[6] “Obituary,” 397-398.

[7] “The Royal Society,” Morning Post (London), April 23, 1874, p. 2.

[8] W. H. Barlow, “The Logograph, or Writing by the Voice,” Popular Science Review 13 (1874): 278-288.

[9] Barlow, “On the Pneumatic Action,” 278.

[10] “Logograph,” Journal, 67.

[11] The rumor can be traced back to Louis Hertz, Antique Collecting for Men (New York: Hawthorn Books, 1969), 269-270, where it is already characterized as a “legend.”

[12] For a facsimile of the traces themselves, see http://vlp.mpiwg-berlin.mpg.de/library/data/lit19823/index_html?pn=14&ws=2, accessed March 30, 2014.

[13] Chas. A. Morey, “The Phonautograph,” American Journal of Science and Arts 8 (Aug. 1874): 130-31.

[14] Specifically, I’ve filled in the area under the curve with white and then run the resulting images through additive synthesis software (AudioPaint) using a 0-22050 Hz linear frequency scale and a white noise sample run through a narrow band filter.

[15] One second reportedly corresponds in this case to six syllables, and measuring each combination of six syllables in both records of “Peter Piper” yields an average width of 1955.5 pixels per second, a figure which I applied in turn to all the records in Barlow’s 1874 paper.  Then, comparing the width of the records of the phrase “recording instrument” as given in his 1874 paper for the Royal Society and his 1878 paper for the Royal Dublin Society, I concluded that the scale ratio of the one to the other was approximately 1 : 1.75, and set the speed of the records in the later article on that basis.  In fact, Barlow appears to have varied the recording speed of his apparatus—see the reference to “different speeds” in W. H. Barlow, “On the Articulation of the Human Voice, As Illustrated by the Logograph,” Scientific Proceedings of the Royal Dublin Society n.s. 2 (May 1879): 153-172, at 169—but I only have the one semi-objective speed reference, which I’ve tried to apply consistently throughout.

[16] Barlow, “On the Pneumatic Action,” 277-86.

[17] Catalogue of the Special Loan Collection of Scientific Apparatus at the South Kensington Museum, Second Edition (London: George E. Eyre and William Spottiswoode, 1876), 160, item 731a, which notes that they also had some “records.”

[18] Karl Pearson, The Life, Letters and Labours of Francis Galton, 3 vols. (Cambridge University Press, 1914-30), 2:192.

[19] “Logograph,” Journal, 65-68.

[20] Barlow, “On the Articulation”; on Barrett’s reading of the paper, see “Royal Dublin Society,” Freeman’s Journal and Daily Commercial Advertiser (Dublin), March 19, 1878, p. 6.

[21] “Logograph,” Journal, 66.

[22] Barlow, “On the Articulation,” 154.

[23] Barlow, “On the Articulation,” 161.

[24] “Logograph,” Journal, 67.

[25] Barlow, “On the Articulation,” 162.

[26] Silvanus P. Thompson, “A New Phonautograph,” Proceedings of the Bristol Naturalists’ Society n.s. 3 (1879-1882): 114-117, at 115.

[27] William Henry Preece, “The Telephone,” Notices of the Proceedings of the Meetings of the Members of the Royal Institution of Great Britain 8 (1875-1878), 501-507, at 502; record shown on page 503.

[28] W. H. Preece, “The Phonograph,” Journal of the Society of Arts 26 (1878): 534-538, at 536.

[29] Sir Frederick Bramwell, “Telephones,” in The Practical Applications of Electricity (London: Institution of Civil Engineers, 1884), 26-55, at 28-30.

[30] Edward Henry Knight, Knight’s American Mechanical Dictionary (Cambridge: Riverside Press, 1880), 2514.

[31] John G. McKendrick, “Experimental Phonetics,” Scientific American Supplement 1380, June 14, 1902, pp. 22112-22115, at 22113.

[32] Frank Bolton and W. H. Preece, “Some Recent Developments of Applied Science, Illustrated by the Phonograph and Other Apparatus,” Royal Engineer Institute Occasional Papers 2 (1878): 203-217, at 214-215.

[33] Chris Carola, “Soundtrack to history: 1878 Edison audio unveiled,” Associated Press, October 25, 2012, http://news.yahoo.com/soundtrack-history-1878-edison-audio-unveiled-064429849–finance.html, accessed March 31, 2014.

[34] Patrick Feaster, Pictures of Sound: One Thousand Years of Educed Audio: 980-1980 (Atlanta, Georgia: Dust-to-Digital, 2012), track 20, pages 103-105.

[35] Barlow, “On the Articulation,” 154.

[36] Barlow, “On the Articulation,” 169.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.