The sound of a bad word has nothing whatever to do with its being bad. Virtually any bad English word will exist in some language somewhere as a good word. We should should also remember that Carlin's two-way words also provide evidence for that. One can say in polite company, "Would you hold Peter?" but not "Would you hold my peter?"
Is it the subject matter, then, that makes a bad word bad? Not entirely, for sure, for we have euphemisms for bad words -- "poop" for "shit" or "take a leakk" for "piss" or "make love" for "fuck" and this suggests that the subject matter per se isn't the problem either. However the social contexts in which these euphemisms occur has a bearing on their acceptability. A mother can ask of her child, "Do you need to make "pee-pee"?" in a gathering of other mothers. However, I don't think a middle-aged middle-class woman would likely say, "Please excuse me. I need to pee," to a group of middle-class women.It appears then that the sound of a bad word has nothing to do with its being bad and that while the subject matter provides the source for bad words we have bad words with unproblematic sources (religion being the best example). The inference I draw is that it is the the word itself that is the problem and its boundaries of use are socially prescribed. A guy can say, "I need to piss," to a drinking buddy in a bar but not to a group of people in a formal business meeting.
In addition to the social context, the culture one grows up in has a bearing on what is and is not a bad word and how bad a bad word will be. In 1995, I taught in England and while wearing a fanny pack (a somewhat outmoded small bag one straps around one's waist) and made the mistake of referring to it by that name. That got quite a response from my students. It seems that "fanny" in British English is equivalent to "pussy" in American English. That example provides further evidence that the sound of a word has nothing whatever to do with its being a bad word.Once while a student at Rice in Houston I found myself in a small gathering of socially prominent people and a middle-agedwoman referred to a group of people as "coonasses." I, who am hard to shock, was shocked. I worried that the "coon" part might be a derogatory reference to African Americans and "ass" speaks for itself. That I was in very polite society and an adult woman had said it contributed to my distress. I later found out that "coonass" is just a slang word for cajuns, though not, I suspect, a totally nonpejorative word.
Not long after I moved to Ohio State, a lawyer in the Law School asked me to testify on behalf of a student who had been arrested for the use of "prurient" language (" Arousing or appealing to an inordinate interest in sex" -- Answers.com), which was against the law back then. Our late unlamented Vice President to President Nixon, Spiro Agnew, was in town and a group of Ohio State students was marching up High Street to downtown to protest his being there. As they marched a "street preacher" kept harassing them ultimately prompting one student to say, "Look, motherfucker, how come there's so many of us and so few of you?" A policeman heard this and arrested him on the spot. Later in court, he testified under oath that when he heard that, he instantly thought of someone fucking his mother. That would make the language prurient but it also made the cop either a liar or someone desperately in need of counseling.I testified that, first, the cop's response was odd on the grounds that "motherfucker" doesn't mean "someone who fucks a/his mother." If it did, sentence (1) would be self-contradictory and it isn't.
(1) Look, motherfucker, I know you don't fuck your mother.
I also testified that in the case of many "fixed" or "frozen" expressions consisting of two or more morphemes, we do not normally "parse" them into their bits. When Barry Goldwater ran for President I didn't think of either gold or water and certainly didn't think of piss when I heard his name. And when I go down to our greenhouse, I don't think of it either as green or as a house. The same goes for "motherfucker." The force it has as an insult does not require that. Nor does the force of "cocksucker" depend on one parsing it into someone who sucks cocks. Another argument I made is that in different subcultures, a given bad word will vary in its force. Back in the 70's when I testified, I noted that "motherfucker" was very commonly used in the African American culture and I might (certainly could) have argued that among young White males, it was also commonly used. The high frequency of use in such subcultures leads to its loss of force. So, for the jury, which consisted of six middle-to-old aged White people, to convict a kid who lives in a subculture in which "motherfucker" is a commonplace curse word because it is shocking to them would be wrong.
Naturally, the jury convicted the kid. Fortunately, the Ohio Supreme Court reversed the conviction on the grounds I had argued at trial.
Three last notes. First, if you learn a language as an adult, it is very unlikely that you will be shocked by any of that language's bad words. Given this fact and the fact that the sound of a word has nothing to do per se with its being a bad word and that even the subject matter or meaning are insufficient to make a word bad, I think we must conclude that what any bad word is is a word our mamas and papas told us was bad. We are Pavlov's dog when it comes to how we react to bad words. Which is to say that we are still primitive people for we still respond to word magic. Second, I have never heard of a language that does not have bad words that are available for use to curse others, so there must be something very right about them. Finally, the more abusive and unacceptable a word is, the more likely it will be used, at least by American men, as a sign of affection for another man. That's the aspect of bad words I really like.