Do People and/or Robots Speak on Pitch and in Harmony with themselves?

Question

I asked this in Linguistics, but apparently nobody there wants to venture an answer; maybe this is a better place to ask it.

My ear is not good enough to tell via observation whether people speak in musical scales, which would seem likely for those with a good ear (talented musicians, and those few who naturally have perfect pitch).

For example, do people normally speak with a certain pitch, say a middle C, and then when they lower or raise the pitch of their voice for modulation, speak in other notes within the C major scale (or C minor, or what have you)?

If they do, do other people find these speakers' voices more pleasant than someone who, for example, normally speaks dissonantly, in a note between C and C#/Db and then uses non-harmonious pitches when varying the pitch of their voice for effect?

On a related note (no pun intended), are the fake/robot voices (Siri and such) calibrated to speak "on pitch" and in harmony with themselves?

Finally, in a more Sherlock Holmesy-type vein, would it be possible to guess a person's favorite type of music based on which scales they use in speaking (major scales = rock or pop, minor scales = blues, 7th or 9th = funk, exotic scales = jazz or classical, etc.)?

Answer 1

There are some facts in this question, but they get tangled in some misunderstandings and assumptions. Let's see if we can sort them out:

• All sound has frequency—you can measure how far apart the sound waves are.
• If one frequency is prevalent enough, we say that the sound has pitch. A tuning fork is pitched, a burst of radio static isn't. A person singing "la" on a particular note is pitched; a person saying "sh" isn't. Frequency is a scientific thing; pitch is a human idea layered on top of it.
• Human speech includes both unpitched noises (consonant sounds) and voiced sounds, like vowels, that might rest on a certain pitch. Steve Reich's Different Trains samples short clips of speech and "maps" the pitches, so they can be imitated by instruments:

• But he manages this by carefully selecting certain moments. Much of your question is built around connecting a person's speech to a single note. Unless the person speaks in a monotone, not unlike a 1960s "Lost in Space" robot voice, the inflection of speech usually moves around rapidly (within a syllable, for instance), and over a fairly broad range. And here's the first problem with analyzing it musically: one of the key differences between things we perceive as pitched and unpitched is that there must be "enough of" a given frequency for us to perceive a pitch. Human speech seldom "stays put" long enough for us to map it to a melody, Different-Trains-style, let alone a single note.
• Other misunderstandings: You talk about speaking "in a scale," or "in harmony." The thing about chords, scales, and similar constructs is that they leave giant gaps in the frequency spectrum. "A" is 440Hz, "B" is 493.88. But speech inflection typically slides through frequencies—the human voice has a very difficult time jumping directly from pitch to pitch; we do better when we interrupt the voiced tones with consonants or pauses. Yes, two different people have a different vocal range, and you could probably identify a "central range" that they spend most of their time in. But no one speaks "in a scale" or a chord.
• That means that the next conjecture, that one person would find another's speech unpleasant for reasons connected to the music-theory notion of "dissonance," doesn't work. It is true that people often find others' speech unpleasant, and the reasons given often have to do with frequency or timbre (e.g. an upstate New York accent described as "whiny" or "nasal"). And some research has been done about how we respond to human voices at certain frequencies (e.g. being irritated, or having parenting instincts triggered, by the frequency range of babies' voices). But there hasn't been much, and it would be pretty hard to draw conclusions that separated conditioning and culture from the results. In particular, when it comes to regional speech patterns, the fact that people within that region don't (usually) perceive it as annoying suggests that it's nothing more than culture shock, and has more to do with the listener being unfamiliar with these vocal patterns than anything explicable by music theory.
• Re artificial human speech, like Alexa and Siri: I don't have documentation, but I would assume surely a lot of work and thought went into selecting ranges and patterns that are positively received. One of the problems plaguing synthesized speech since its inception has been "that computer voice is annoying." Advancements have had a lot to do with bringing inflection to the Lost-in-Space monotone, and have perhaps skirted the Uncanny Valley in attempts to get it right. It's notable that by now you can choose several "Siris" with regional accents and male or female registers: What people find most acceptable varies personally and regionally. Again, though, the only thing this has to do with music is the Venn diagram in which both speech and music use sound, which has frequency. People might describe a pleasing voice as "musical," and that might mean that its inflection varies more, or that its voiced sounds are more pitched (as opposed to a "rasping" or "breathy" voice in which frequencies would be more diffuse and unpitched), but that doesn't mean that musical conventions (especially ones invented to explain the practices of a fairly narrow slice of human history, modern Western music) can be projected onto it with any meaningful results.

Answer 2

The tonal language into dialect of spoken Mandarin Chinese involves smooth portamentos/glissandos/bends when speaking word tones, such as rising (e.g. Mandarin for "white") or descending (e.g. Mandarin for "down"). Only singing exact notes in a diatonic scale, such as the C Major scale, instead of using those pitch bends when speaking those word tones would sound strange and Auto-Tune-like.

I've read studies that indicate that Chinese speakers use shockingly consistent notes (absolute pitch-wise) when speaking the same word over different days, but no indication of how closely those notes match 12TET A440(Hz).

(While the Chinese do also love their sung music, I find understanding those Chinese lyrics to be hampered compared to understanding spoken Chinese due to word tones typically being destroyed in the process. For example, the first word of the Chinese folk song "Mo Li Hua" is spoken with a dipping-and-rising tone in Mandarin but sung completely flat.)

Answer 3

Nope! People will not conform to scales in general speech. Check out this video, where a musician Youtuber (with an excellent musical ear) tries to play his piano in harmony with a video of a pastor asking for money. Some of the things the pastor says do precisely fall on notes of a scale. But most don't. (And the Youtuber humorously tries to get it to sound good anyway)

Couple of things to consider... We are used to standard tuning, where an A4 note is 440Hz. But the existence of a ubiquitous tuning system is not, well, ubiquitous. Depending on the culture or time period you're from, it may be that music is not tuned to a common tuning. So the same scale (e.g. C major) would sound at different pitches in different tuning systems. Tuning systems are a human construct, and have no basis in physiology of the human voice.

Also worth considering, most people aren't musicians, and most musicians aren't singers. Ask a random person to sing, and they won't be able to sing on key. They will tend to sing notes that "don't exist". If the note is a C, they may sing that C a bit sharp, but not quite sharp enough to be an actual C sharp... rather it's somewhere in between the two notes. Many humans can "match pitch" of another human's voice, but most can't match pitch of an instrument that is tuned to standard tuning.

Answer 4

When pitch is used in language it is just enough to either voice and not voice the sounds of speech or to give a relative direction to a pitch bend.

Music on the other hand treats pitch very precisely for the fundamental elements of scales and chords.

One language speaker may use a lot of pitch while another might speak comparatively monotone as a matter of personal style. Musician have no such personal style choice when it comes to the pitch of scales and chords where you are either in tune or out of tune.

Some languages, like Chinese, use pitch as an important element of speech, but I don't think it goes beyond just enough pitch to perceive either upward or downward pitch bends.

Music does have personal style choice with some aspects of pitch like vibrato, but those are separate from the fundamental elements of scales and chords.

For example, do people normally speak with a certain pitch, say a middle C...

In music there is something called a recitation tone, but that is singing.

...speak in other notes within the C major scale (or C minor, or what have you)? ... which scales they use in speaking (major scales = rock or pop, minor scales = blues, 7th or 9th = funk, exotic scales = jazz or classical, etc.)?

This too would just be called singing.

..someone who, for example, normally speaks dissonantly, in a note between C and C#/Db and then uses non-harmonious pitches when varying the pitch of their voice for effect?

You are mixing musical and linguistic concepts. Dissonance and lettered pitches are musical concepts. People might use differing amounts of pitch in their speech, but there isn't any concept of speaking dissonantly or speaking out of tune. Not in the sense of what dissonance and out of tune mean in music.

You seem to be playing with the idea of a language speaker increasing the precision of their pitch enough to make their "spoken" pitches match scales. That just sounds like a basic definition of singing. Although, if you equate singing with our traditional conception of a song, you might think singing is something else. You might look into things like chant, work songs, and street cries, which is where functional speech, rather than poetic song lyrics, are given pitch, and you might think of those forms as a prototype of singing and song. They are essentially your idea of people speaking on pitch and in harmony (harmony in the sense of the pitches fit into some kind of tonality.)

Answer 5

I can't imagine any language spoken solely on the notes of the scale. Speech isn't in tune!

In English speech, the pitch of the voice is constantly changing: from syllable to syllable and even within a syllable. The word "Why?", for example, may be pitched higher or lower depending on our concern, curiosity or indignation, and it might slide upwards, downwards, or downwards-then-upwards. These inflections make up our tone of voice. We also use a higher pitch to emphasize a particular syllable or to change the meaning of a word: SUR-vey (noun) sur-VEY (verb).

Although we sometimes come close to singing, for example when calling a lost dog, speaking entirely on the notes of scales is singing.

By the way, the note between C and C sharp is only ever dissonant in the context of music: not in the context of speech. (Even in the context of music it needn't be dissonant: if, say, the whole piano is tuned a quarter-tone flat.)

We tend to find a speaking-voice monotonous if it uses a narrow vocal-range. We like variety. When we describe someone's speaking-voice as 'musical' we mean only that it visits a fairly wide range of pitches.

Um.

Sprechgesang - formulated by the composer Arnold Schoenberg - is a mixture of music and speech, the singer sliding from one pitch to the next. (You can hear it in Pierrot Lunaire.) Janacek incorporated the precise rhythms and inflections of everyday speech (in Serbo-Croat!) into even his instrumental music. His woodwind sextet, for example, begins with the oboe playing the word "Mladi" (youth), which is also the title of the piece.

Answer 6

I can't imagine, that for most parts of speech one is able to determine a frequency at all, due to the high (dependent on the language) proportion of consonants, which in the musical sense are more noise than tone.

On the other hand, the amount of movements needed with tongue, muscles in the mouth region, palate and so on, which is necessary to get the pronounciation right is considerable. All this movements would change the pitch if I would play a woodwind instrument instead of speaking and the pitch would change smoothly and not with the strict steps required for a simple scale.

And finally considering the hard work professional singers have to do to get the words into the correct sound/pitch for an aria - I don't believe, there is a chance for this to happen accidentally.

Answer 7

There must be cultural and personal variations of preference of intervals in speech. The reason I think this is that in one job, I had a whole bunch of Polish co-workers. SOME, but not all of the women, had a certain interval between words that they favored frequently. This even carried over to when they were speaking other languages than Polish. To my ear, the interval sounded kind of dissonant, which gave a kind of skeptical or judgmental flavor to the content.

Second example: in some cultures, the woman's voice, in order to be considered polite, must be pitched higher than their natural speaking voice.

In the US, what is considered an attractive voice (as opposed to, for example, one that might be considered shrill) has shifted somewhat lower over the last couple of generations.

There is actually a way to determine what an individual's natural speaking voice range is.

Note that some conditions listed in the DSM sometimes affect an individual's speaking voice range, pushing it higher than natural, lower than natural, or a narrower band (i.e. the voice is considered "flat" or fairly monotone.

I realize this isn't precisely what you asked, but I believe this type of observation will get you closest to the general topic you are curious about.