And I am glad of it with all my heart:
I had rather be a kitten and cry mew
Than one of these same metre ballad-mongers;
I had rather hear a brazen canstick turn'd,
Or a dry wheel grate on the axle-tree;
And that would set my teeth nothing on edge,
Nothing so much as mincing poetry:
'Tis like the forced gait of a shuffling nag.
— Hotspur, Henry IV, Part I, Act III, William Shakespeare
Bad poetry is up there, but it can't compare with fingernails on a chalkboard for sheer obnoxiousness. I've seen various explanations — such as how the sound is similar to that of a child in distress or the cry of a macacque monkey which preyed on our primate ancestors — but none of them were particularly convincing. Then I came across this one.
In "Psychoacoustics of a Chilling Sound," the authors describe their study using two dozen adults served as guinea pigs for rating sounds on a pleasantness scale. My personal favorite has to be how they created a standard — much like how there are standards for the meter, liter, second, etc. against which unknown quantities are compared — for the sound of fingernails on a chalkboard. (Just thinking about it makes me cringe.) They constructed a "True Value Pacemaker" using a three-pronged garden tool dragged over a slate surface. (Aieeeeee!)
Now that's what they should have used to get Noriega out of that church he was holed up in, instead of Barry Manilow. (Then again, I think I'd much rather listen to nails on a chalkboard, or even Lou Reed's Metal Machine Music than Barry Manilow. Hell, I'd even rather listen to Fran Drescher's donkey-bray laugh than to Barry Manilow.) It's really amazing what researchers can get funding for, isn't it? Anyway, here's a summary of the whole issue from the Straight Dope.
In the aforementioned scientific paper (which appeared in a publication sternly entitled Perception & Psychophysics, and is not to be confused with a vulgar and sensationalized, if entertaining, article that appeared subsequently in Psychology Today), the authors note the antiquity of human curiosity on this subject. No less an authority than Aristotle acknowledged the "aversive quality" of scraping sounds. Our heroes even dug up the archaic English verb gride, which means to make godawful noises by means of scraping or cutting.
Getting down to business, Halpern and friends subjected 24 adult volunteers to various noises with a view to determining whether blackboard scraping was really as excruciating as it was made out to be. Generally speaking, they found, it was. (For purposes of reproducibility, the scraping was conducted not with fingernails but with a three-pronged garden tool, solemnly described as a "True Value Pacemaker model.") Interestingly, "rubbing two pieces of styrofoam together," the sound that results when you pry two styrofoam cups apart, came in second.
Next, by means of the magic of high tech, the researchers filtered out the most high-pitched portion of the scraping sound. To their great surprise, what remained was as unpleasant as ever. However, when they filtered out just the lower frequencies (particularly 3.0 to 6.0 kilohertz, for you weens), they found that what was left was relatively bearable--"quaint" or "tinkly," in Blake's description. In other words, it was the low-to-middle frequencies, not the high ones, that really set people's nerves on edge."Why is the sound of fingernails scraping a blackboard so annoying?," The Straight Dope
I tried to track down the original Halpern paper online, but didn't have any luck. It was, after all, published in 1986 when dinosaurs walked the earth. (Perception & Psychophysics doesn't have issues this old online yet.) I did, however, find a nice summary, even if it is one giant paragraph (I guess whitespace costs more in the Netherlands. Must be some Euro or VAT thing.):
Halpern et al. (1986) examined the unpleasantness of a chilling sound. Although this study investigated the unpleasantness of a sound instead of the human ability to perceive properties of the sound source, it is included because of the very similarmethod and unexpected results. In a first experiment, subjects had to judge the unpleasantness for a number of different sounds, such as jingling keys, a blender motor, and scraping metal. The sounds were matched in duration (3 s) and amplitude (equal maximum value). The results showed agreement between subjects regarding the unpleasantness of the sounds. The sound judged to be most unpleasant was that produced by slowly scraping a three-pronged garden tool over a slate surface, a sound very similar to the sound of fingernails scratching across a blackboard. The spectrogram of this chilling sound revealed several prominent harmonics, the lowest at 2.8 kHz. The amplitude waveform showed an aperiodic temporal structure with a rapidly fluctuating amplitude envelope. To investigate the contribution of spectral content to the sound’s unpleasant character, the authors removed energy from different frequency regions by either highpass or lowpass filtering. The sounds were matched in amplitude by equalizing their RMS value. Subjects had to rate the unpleasantness of the filtered sounds and were told how the stimuli were created before listening. Results showed that decreasing the lowpass filter cutoff frequency from 8 to 3 kHz had no effect on the unpleasantness ratings. Increasing the highpass filter cutoff frequency from 2 to 6 kHz, the sound lost some of its unpleasantness, with a large drop in unpleasantness between 3 and 4 kHz. Apparently, removal of lower frequencies, not of the highest ones, lessened the sound’s unpleasantness. In this experiment the sounds were matched by their level, but still may be perceived as not equally loud. A third experiment tested the possibility that unpleasantness had been confounded with loudness. Subjects listened to a selection of stimuli from the previous experiment, presented at two sound pressure levels 10 dB apart, and had to judge the loudness. An intensity decrease of 10 dB resulted in an estimated loudness drop between 41% and 50%, confirming that subjects were estimating the loudness. Sounds presented at the same sound pressure level showed no difference in the estimated loudness, indicating that loudness differences could not have influenced the unpleasantness ratings. In a final experiment, the contribution of temporal fine structure was evaluated by presenting subjects with four different stimuli: the original sound, a demodulated version of the original (the original sound divided by its temporal envelope contour), an unmodulated synthesized sound (sum of three sinusoids corresponding to the first three prominent harmonics of the original sound), and a modulated synthesized sound (the sum of three sinusoids multiplied by the temporal envelope contour of the original sound). The subjects’ unpleasantness ratings of the original sounds were much higher than those of the synthesized sounds, indicating that the latter did not mimic the original chilling sound very well. No differences were found between the original and demodulated original sounds and between the unmodulated synthesized and modulated synthesized sounds, indicating that temporal envelope structure did not contribute to the unpleasantness of the sounds. It is still unclear why this sound is so unpleasant for human listeners. The authors wonder “whether it mimics some naturally occurring, innately aversive event” (p. 80), and think of warning cries or vocalizations of some predator. But, “regardless of this auditory event’s original functional significance, the human brain obviously still registers a strong vestigial response to this chilling sound” (p. 80)."The sound of rolling objects: Perception of size and speed" by Mark Mathieu and Jeanny Houben
The ranking for sounds from most pleasant to least pleasant is rather intriguing:
- Spinning Bicycle Tire
- Running Water
- Jingling Keys
- Pure Tone
- Pencil Sharpener
- Shaking Metal Parts
- White Noise
- Compressed Air
- Blender Motor
- Dragged Stool
- Metal Drawer Opening
- Scraping Wood
- Scraping Metal
- Rubbing Styrofoam Pieces Together
- Scraping Slate with Garden Tool (fingernail/chalkboard)
I never found styrofoam rubbing together to be particularly grating, no pun intended, though. Conspicuously omitted from the list, however, is the sound of MTA subway screeching. Now that's a sound that just about rips out one's spinal column and skull ala Predator.
Anyway, the interesting observation is that application of a low-frequency filter drops the annoyance factor measurably, demonstrating that lower frequencies are more annoying than higher ones. (I always found the higher tones in Fran Drescher's voice to be the most annoying, but I don't get paid to do psychoacoustic research, so what do I know?) I guess someone should ask William Tager about annoying frequencies. (Hint: he's the one who attacked Dan Rather for beaming thought waves at him, and this inspired "What's the Frequency Kenneth" from REM.)
But why should lower frequences be more annoying than higher ones? The answer seems to lie in the physiology of the human ear:
Having recently done some work on the pleasantness/unpleasantness of sounds (JASA 110(1), 380-390, 2001) I was somewhat curious to read that it was the low frequencies that produced the effect, since most of the literature I am aware of (e.g. see review in Vitz (1973) P&P, 11, 84-88) suggests the opposite. However, having got hold of a copy of the Halpern et al. paper I note that sound in question has a fairly strong harmonic structure with a fundamental at about 1.4 kHz. The fundamental is very weak and most of the energy is in harmonics 2,3,4 and 5, starting at 2.8 kHz. By most standards this sound would be considered to be quite high. So the term "low" should be considered in relative terms. Nevertheless, application of a high-pass filter to this sound suggested that it was frequencies less than 2-3 kHz which were predominant in the effect, and by implication the fundamental and possibly the 2nd harmonic, i.e. sounds between about 1 - 2 kHz. In the previous literature and my own work, sounds less than 1 kHz were considered to be least annoying or most pleasant. So why do sounds with frequencies between 1-2 kHz cause the effect? My guess is that the effect is produced by activation of various myogenic reflex responses including the stapedius response, the post-auricular response and responses of other muscles innervated by the facial nerve (and possibly the trigeminal nerve). It so happens that the tuning curves of stapedius motorneurons have their best frequencies between 1-2 kHz with a threshold of about 75 dB in the cat (see Kobler et al. (1992), J. Neurophysiol. 68, 807-817). (These should be distinguished from myogenic vestibular responses mediated by the accessory nerve, which responds to frequencies less than 1 kHz.) In order for this to work then the scraping sound would have to be above about 75 dB, but it's not clear from Halpern et al. what intensity they presented the sounds to the subjects. However, the proposed mechanism would account for why the effect appears to be reflexive. It can't be very pleasant having all those muscles twitching away!"Re: finger nails on blackboard" by Neil Todd, todd(at)FS4 dot PSY dot MAN dot AC dot UK
Ok, let's put that into English. The stapedius is a muscle in the inner ear that acts to protect the ear from loud noise, including the sound of our own voice (Fox newscasters must have an overactive one to stave off deafness) and mastication (eating). Certain sounds in the 1-2KHz range have the effect of causing spasms in the stapedius. The exact mechanism is unknown, but may have to do with higher frequency harmonics arising out of the lower tones, perhaps because of the resonant frequency of the ear bones. The effect of certain frequencies is to cause pain in the stapedius which makes us cringe.
So how plausible is this? Well, consider that the distaste for screeching does not seem to be be universal among primates, which it really should be if the mechanism for distaste is one of avoidance of a predator. Tamarin monkeys, for example, don't seem to mind it much:
As a second test of whether tamarins might have acoustic preferences based on something other than amplitude or behavioral relevance, we attempted to generate two nonmusical stimuli with similar amplitudes that were expected to produce a large preference in humans. We began by generating a stimulus that is highly aversive to most humans—the sound of fingernails on a blackboard (Halpern, Blake, & Hillenbrand, 1986). The relationship between the responses that humans have to this stimulus and to musical stimuli is unclear, but it seemed conceivable that nonhuman animals might respond aversively to such a stimulus despite the lack of preference for consonance over dissonance.
When tested on the screech and control stimuli, however, the tamarins showed no evidence of a preference. We ran the tamarins for several consecutive sessions (NZ37 sessions) to see if a preference would emerge over time. As shown in Fig. 5b, there was no preference (tZ0.89; pZ0.15). In contrast with humans, who show a pronounced preference for white noise over the screeching sound, tamarins do not exhibit a preference."Are consonant intervals music to their ears? Spontaneous acoustic preferences in a nonhuman primate" by Josh McDermotta and Marc Hauserb, Cognition, 94 (2004)
Now, just when you start to believe that it's all in the ear, I'll throw this into the mix:
Seth Horowitz is a neuroscientist who uses Magnetic Resonance Imaging (MRI) scans of the brain to find how different sounds can trigger activity in the brain. Now he is working with a musician to create music incorporating sounds, in the hopes of triggering specific emotional responses.
Horowitz began to study whether sounds could trigger emotional states. For example, the sound of fingernails on a blackboard is an effective way to cause many people to feel uneasy. So Horowitz broke down the sound of fingernails on a blackboard to isolate exactly the sounds responsible for triggering uneasy feelings. He calls those sounds neuro-sensory algorithms, or NSAs. Then he analyzed sounds that trigger activity in the same region of the brain. He did the same for sounds that make people feel calm or happy or stimulated. By combing through the data, Horowitz was able to come up with dozens of different sounds that triggered emotional responses in the correct regions of the brain. NSAs all sound different. Some can be a very brief sound that immediately triggers activity in a certain part of the brain. Others can be complicated mixes of sound that last up to 30 seconds and trigger activity in different parts of the brain simultaneously."Mood Music" The Osgood File (CBS Radio Network), 26 January 2005
Anyway, the next time some brain-dead creationist tells you that the human body is an example of intelligent design, I suggest you scrape your fingernails across a chalkboard. (Or a piece of slate with a gardening tool.) Get them to explain the "intelligent design" behind that for you.
Sources and Further Reading
- "Psychoacoustics of a chilling sound." by D. Lynn Halpern, Randy Blake, Jim Hillenbrand, Perception and Psychophysics, Vol. 39 No. 2, February 1986, Pages 77-80
- "Why is the sound of fingernails scraping a blackboard so annoying?," The Straight Dope
- "The sound of rolling objects: Perception of size and speed" by Mark Mathieu and Jeanny Houben
- "Re: finger nails on blackboard" by Neil Todd
- "M109: Reflexes and/or associations" (more discussion on Neil Todd's comments)
- "Are consonant intervals music to their ears? Spontaneous acoustic preferences in a nonhuman primate" by Josh McDermotta and Marc Hauserb, Cognition, 94 (2004)
- "Mood Music" The Osgood File (CBS Radio Network), 26 January 2005