13

Daniel Everett claims in Don't Sleep, There Are Snakes (Ch. 11) that the English "p" and "b" in "pin" and "bin" are separate phonemes, since they alone distinguish the words "pin" and "bin," whereas the different "p" sounds in "pin" and "spin" are the same phoneme, because even though they sound different, the difference is never required to distinguish two words.

But isn't the "p" in "spin" really just a "b"? I can't hear a difference between the way I naturally say "spin" and what I say when attempting to pronounce "sbin."

  • 4
    Well, there is the word "hasbeen", which is pronounced "hasbin" in many areas. If you pronounce that, doesn't the b sound different than a p? It does for me. – Mr Lister Dec 1 at 12:05
  • 5
    @MrLister The problem with that is the S in 'hasbeen' represents the voiced sound /z/. – Aeon Akechi Dec 1 at 23:28
  • 1
    It's Don't Sleep, There Are Snakes by the way. I fondly remember reading this account! – Vandermonde Dec 1 at 23:55
  • @Vandermonde Thanks for the correction – WillG Dec 1 at 23:57
  • 3
    Yes, but can you hear a difference between "horse pin" and "horse bin"? Then get rid of the "hor". – Dawood says reinstate Monica Dec 2 at 5:12
35

It is kind of convention to assign the phonemic value /p/ to the p in spin, since there is no minimal pair /p/:/b/ in this environment (words like *sbin don't exist).

Now comes the fun part: In English, there is a double contrast between /p/ and /b/ in initial position: The p is voiceless and aspirated [ph], the b is voiced and not aspirated [b]. The p in spin is in-between: voiceless and not aspirated [p]. So, is it a de-aspirated [ph] or a devoiced [b]? Conventionally, the first choice is assumed. But I have seen German dialect orthographies where <schb> was choosen as the grapheme for [ʃp] preferring the second type of analysis.

  • 1
    Is word-initial /b/ really voiced? I remember when learning French (as a native AusEng speaker) that I had to learn to pronounce /b/, /d/ and /g/ more "heavily" (ie, more voiced) than I would naturally do. (I can barely hear any difference between the way I pronounce French pas and English bar) – Steve Bennett Dec 2 at 3:00
  • 1
    @SteveBennett and how do you differentiate between bar and par? My language background is German but I think there is an audible different sound when speaking both words – eagle275 Dec 2 at 10:57
  • 2
    "words like *sbin don't exist" Unix and Linux engineers beg to differ. See: google.com/search?q=usr+local+sbin Also… "Don't try to bend the sboon with your mind. That is impossible. Try to realize there is no sboon." – Bruno Bronosky Dec 2 at 21:16
  • 1
    @eagle275 the "p" in par is aspirated, so they're pretty different. Now that I think about it, I struggle with some American accents that don't really aspirate t's. Every time I hear Ira Glass say "Lagunitas" I hear "Lagunidas". – Steve Bennett Dec 2 at 23:03
  • 3
    @BrunoBronosky: How do you pronounce sbin in that context? I always say es-bin. – jknappen - Reinstate Monica Dec 3 at 11:02
15

First, there is a lot of variation in English, so don't expect the facts to be the same for all speakers. Second, it's unclear what you mean by "really". There is phonological analysis, and there is acoustic analysis. The standard mostly-phonological analysis is that "pin" has aspirated [pʰ] and "spin" has unaspirated [p], but they reflect a single phoneme /p/, which contrasts with /b/. It is often said that that /b/ is voiced, but for most speakers it is not always voiced in the way it is in French, so there is an alternative analysis of the phonemic contrast /p,b/ as being one of aspiration (the phoneme "p" is treated as aspirated and "b" is voiceless and unaspirated).

If we say that "p" is /pʰ/ and "b" is /p/, the description of the allophonic difference between the variants of "p" is harder to explain. One solution is to invoke an additional property, tense / lax or fortis / lenis, so /p/ is fortis and /b/ is lenis, and then you are free to say that fortis stops can be aspirated in syllable-initial (or foot-initial) position. Without some additional feature, and if you are not allowed to say that /b/ is voiced (given the presumption that there is no vocal fold vibration during its production), then you need something to distinguish the labial stops of "pit, bit, spit, happy, cabby". Actually, in words like "cabby", "b" is produced with vocal fold vibration. The confusion situation is because people do not agree on the criteria for saying what some language sound "really" is – is it a claim about phonological pattern, or a claim about phonetics.

It may be useful to gather some good recordings of yourself saying "spit", "pit", "bit" and so on, and perform the editing test, where you (use Praat to) remove the fricative from "spit". Then the question is, can you tell that "spit" with [s] removed is distinct from initial "b", or do "(s)pit" and "bit" become indistinguishable. It depends on the task that you set for yourself. If you set it up to ask "does this sound most like pit or bit", you will never think that edited (s)pit sounds like "pit", therefore you will always say "sounds like bit". But if you do an a-b comparison (do these two sound the same), you might distinguish (s)pit from bit. That is, phonetically they do sound different. (You have to be very dedicated to the question, in order to be willing to put in the work necessary to do this experiment). Alternatively, you could just make many recordings of "spit" and "bit" and measure voice onset time to see if there is any difference in VOT between /p/ and /b/ in this context.

  • voiced b and non-voiced b entirely depends on the following letter .. if a vocal follows .. its usually a voiced b - but at a word-end or with following non-vocal it "glides" over to a p-like sound (except double b like in abbot ..) – eagle275 Dec 2 at 11:00
4

In English, aspirated "p" as in "pin" ([pʰ]) and unaspirated "p" as in "spin" ([p]) are allophones: two different phones that represent the same phoneme /p/.

However, there are languages that do make a phonemic distinction between the two, so rather than being two allophones of a single phoneme, they are separate phonemes /p/ and /pʰ/ in their own right. For example, in Hindi, they form a minimal pair that provides the distinction between "palā" (a form of the verb "to rear") and "phalā" (a form of the verb "to bear fruit"). (https://courses.washington.edu/hindi31x/101/minimal-pairs.html)

You don't hear a significant difference between "spin" and "sbin" because English also does not have a consonant cluster "sb", nor does it have separate allophones [bʰ] and [b] for /b/.

  • 2
    The question is: how do you know the [p] in "spin" doesn't represent the phoneme /b/? If indeed it isn't aspirated. – reinierpost Dec 2 at 10:35
  • except for names ... Hasbro ... has your sb - and the b is voiced ^^ – eagle275 Dec 3 at 7:41
  • 1
    @eagle275 surely that's /z/ though – Chris H Dec 3 at 10:33
  • 1
    That's not a consonant cluster. That's a word with two syllables "Has-" and "-bro", not "Hasb-"/"-ro" or "Ha-"/"sbro". – chepner Dec 3 at 12:42
  • @Beanluc It's a legitimate question, though one that (I believe) would be resolved by doing a frequency analysis of the pronunciation of "spin", which would show that the consonant following "s" is closer to /p/ than /b/. – chepner Dec 3 at 12:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.