Harlem Shake: A linguistic perspective
So, Internets, you have a new favorite meme the last week or two - Harlem Shake. I'm sure you've all seen one or more videos of this. As someone who's into language things, I found the repeated non-English line interesting and will say a few words about it.
When I first watched a Harlem Shake video, I didn't know anything about the song, the dance, or the meme, so when I heard the non-English line it puzzled me. Having no context, this is what I thought I heard (a very broad transcription):
[ko:lo fred.e.ri:ka]
Basically, I parsed it as two words: something like "kolo", which means something like "wheel" in some Slavic languages, and something like the feminine name "Frederika" (fem. of "Frederick"). Clearly I was missing something, and the brain tries to make sense out of what seems like nonsense. (cf. EVP).
I looked up the song (and discovered the meme was already on Wikipedia). It turned out that the phrase was supposed to be Spanish, something from a poem, to be transcribed in literary fashion as "con los terroristas".
How on earth do you get "kolo Frederika" out of "con los terroristas"?
Actually, it's quite easy.
The most important clue is the fact that in some Latin American Spanish dialects, [s] -> [h]. So instead of
[ko:n lo:s ter.ro.ri:s.ta:s]
you would get
[ko:n lo:(h) ter.ro.ri:(h).ta:(h)]
in which all the [s] became [h] or omitted altogether.
But how does "con los" become "kolo" and "terroristas" become "Frederika"?
First, you have to know that within the mouth, [n] shares a location (place of articulation) with many other consonants, including [l]. So it's quite easy for the two sounds to either blend together entirely or for your ear to mistake what it hears and interpret it as a single sound or a different sound altogether. A really good example in English is the word "illegal". The "il-" prefix is actually the same as "in-" in words such as "indefinite" aka "not definite". However, there is a word formation rule that is used with this particular "in-" that requires it to mutate. With words (bases) that start with [l], the [n] turns into an [l] as well (or merges with it). This happens with other sounds - the [p] at the beginning of "patient" causes the [n] to turn into [m] for "impatient": it's the same prefix ("in-"). So regardless of whether the performer actually articulates both the [n] and [l] (or not) in "con los", it's quite easy for the ear to mistake it for a single [l]. (There is more than can be said about this, in terms of the quality of the vowel in "con" but I don't want to go that in depth.) So now we know that "con los" can be interpreted as "kolo".
Let's look at "terroristas" -> "frederika" now. There are several things happening here with consonants as well, starting with the [t] -> [f]. Pronunciation (as well as aural interpretation of pronunciation) is often affected by what comes before as well as what comes after. In the contect of "con los" becoming "kolo", the [n] was affected by the sound that comes after - the [l] made the preceding consonant [n] change. Likewise, in "impatient", the [p] affects the preceding [n] and turns it into [m]. However, this can happen in the opposite direction - the first sound can mutate the following sound. In my interpretation of [t] as [f], we have to look at the end sound of "los", which in this performance is [h]. [h] belongs to a different category of consonants than [t]; [h] is more similar to [f] and [th]. When you speak, there are often situations when you (of course unknowingly) mutate the sound you produce because it is "easier" to make a sound similar to what you just made than to make a different sound; this often happens when there is no chance of confusion if you make that sound instead of the "proper" sound. In this case, it is easy to either say or hear [f] instead of [t] because the [t] is preceded by [h]:
[h] is very different from [t]
[h] however has a sibling, [f], which is in the same location as [t]
therefore, [h] encourages [t] to become [f] because that's a shortcut
The change of location of the first [r] is a listener memory error. However, it is a common one, historically speaking, and there are many words in a variety of languages whose [r] in some words have moved from the end of a syllable to the beginning of one and vice versa, or done something even wackier. Think about how you pronounce the word "comfortable". Do you pronounce it [kam.for.ta.buhl]?
What about [r]->[d]? This is also a common listener error, especially for speakers of (American) English because we have wacky [r] sounds compared to many other languages. Some [r] sounds are similar to what is called a 'tap' in phonetics - think of the way people pronounce the word "bottle", especially with, say, a stereotypical Cockney accent. The 'tt' is kind of swallowed, but in other dialects - like general American English - it sounds kind of like [d] (but it isn't actually [d]). That is a 'tap'. In the right context, [r] sounds are mistaken by English speakers for 'tap' sounds. Family anecdote: for decades, my father - who only knows a few words of Croatian, learned from his mother and grandmother - thought that the expression for "great-grandmother" was [sta:da ba:ba] and assumed it was both written and pronounced with a [d]/"d". As a teenager, I looked it up in a dictionary and found out it was actually "stara baba" ("stara" meaning "old" in many Slavic languages). As a child he had interpreted the [r] as [d] or 'tap' and was steadfast in that belief every time he encountered it afterwards.
So then how do we get the ending of the word "frederika" from "terroristas"? Again, quite easily, and probably listener error again. The two "s"'s fall out - "istas" -> "i(h)ta(h)". We are left with the mutation (or listener error) of [t]->[k]. This is also an incredibly common kind of mutation and also goes in the opposite directions. (Hint for those learning Italian - a lot of those "-tt-" patterns in words were originally "-ct-" in Latin: look at "perfect" and "perfetto", for example.) So it's not surprising that when the brain can't figure out a completely unfamiliar or foreign phrase that you think you hear one thing when it's actually another. On later listenings, I heard the [t] instead of the [k], but the first time I swore it was "Frederika" because that was a cluster of sounds (aka a word) that made sense to me, even if it didn't match the rest of it. (Incidently, there is a word "kolo" in the Slavic language I'm learning right now. It means "bicycle" and, in compound nouns, "wheel": cf. kolodvor, literally "wheel hall", translated as an old-fashioned term for "railway station". If you go to Croatia, probably you will see Glavni Kolodvor (lit. "head wheel hall"), the central train station in Zagreb.)
So, to wrap things up, there are completely logical linguistic rules that dictate how you may hear and misinterpret the Spanish phrase that is part of the Harlem Shake meme. The real clue to the whole thing is that [s]->[h] is a common element of some varieties of Latin American Spanish. The phrase seems to come from a performance of a poem by Hector Delgado. Delgado comes from Puerto Rico. And one identifying element of Puerto Rican Spanish? You guessed it, [s]->[h].
When I first watched a Harlem Shake video, I didn't know anything about the song, the dance, or the meme, so when I heard the non-English line it puzzled me. Having no context, this is what I thought I heard (a very broad transcription):
[ko:lo fred.e.ri:ka]
Basically, I parsed it as two words: something like "kolo", which means something like "wheel" in some Slavic languages, and something like the feminine name "Frederika" (fem. of "Frederick"). Clearly I was missing something, and the brain tries to make sense out of what seems like nonsense. (cf. EVP).
I looked up the song (and discovered the meme was already on Wikipedia). It turned out that the phrase was supposed to be Spanish, something from a poem, to be transcribed in literary fashion as "con los terroristas".
How on earth do you get "kolo Frederika" out of "con los terroristas"?
Actually, it's quite easy.
The most important clue is the fact that in some Latin American Spanish dialects, [s] -> [h]. So instead of
[ko:n lo:s ter.ro.ri:s.ta:s]
you would get
[ko:n lo:(h) ter.ro.ri:(h).ta:(h)]
in which all the [s] became [h] or omitted altogether.
But how does "con los" become "kolo" and "terroristas" become "Frederika"?
First, you have to know that within the mouth, [n] shares a location (place of articulation) with many other consonants, including [l]. So it's quite easy for the two sounds to either blend together entirely or for your ear to mistake what it hears and interpret it as a single sound or a different sound altogether. A really good example in English is the word "illegal". The "il-" prefix is actually the same as "in-" in words such as "indefinite" aka "not definite". However, there is a word formation rule that is used with this particular "in-" that requires it to mutate. With words (bases) that start with [l], the [n] turns into an [l] as well (or merges with it). This happens with other sounds - the [p] at the beginning of "patient" causes the [n] to turn into [m] for "impatient": it's the same prefix ("in-"). So regardless of whether the performer actually articulates both the [n] and [l] (or not) in "con los", it's quite easy for the ear to mistake it for a single [l]. (There is more than can be said about this, in terms of the quality of the vowel in "con" but I don't want to go that in depth.) So now we know that "con los" can be interpreted as "kolo".
Let's look at "terroristas" -> "frederika" now. There are several things happening here with consonants as well, starting with the [t] -> [f]. Pronunciation (as well as aural interpretation of pronunciation) is often affected by what comes before as well as what comes after. In the contect of "con los" becoming "kolo", the [n] was affected by the sound that comes after - the [l] made the preceding consonant [n] change. Likewise, in "impatient", the [p] affects the preceding [n] and turns it into [m]. However, this can happen in the opposite direction - the first sound can mutate the following sound. In my interpretation of [t] as [f], we have to look at the end sound of "los", which in this performance is [h]. [h] belongs to a different category of consonants than [t]; [h] is more similar to [f] and [th]. When you speak, there are often situations when you (of course unknowingly) mutate the sound you produce because it is "easier" to make a sound similar to what you just made than to make a different sound; this often happens when there is no chance of confusion if you make that sound instead of the "proper" sound. In this case, it is easy to either say or hear [f] instead of [t] because the [t] is preceded by [h]:
[h] is very different from [t]
[h] however has a sibling, [f], which is in the same location as [t]
therefore, [h] encourages [t] to become [f] because that's a shortcut
The change of location of the first [r] is a listener memory error. However, it is a common one, historically speaking, and there are many words in a variety of languages whose [r] in some words have moved from the end of a syllable to the beginning of one and vice versa, or done something even wackier. Think about how you pronounce the word "comfortable". Do you pronounce it [kam.for.ta.buhl]?
What about [r]->[d]? This is also a common listener error, especially for speakers of (American) English because we have wacky [r] sounds compared to many other languages. Some [r] sounds are similar to what is called a 'tap' in phonetics - think of the way people pronounce the word "bottle", especially with, say, a stereotypical Cockney accent. The 'tt' is kind of swallowed, but in other dialects - like general American English - it sounds kind of like [d] (but it isn't actually [d]). That is a 'tap'. In the right context, [r] sounds are mistaken by English speakers for 'tap' sounds. Family anecdote: for decades, my father - who only knows a few words of Croatian, learned from his mother and grandmother - thought that the expression for "great-grandmother" was [sta:da ba:ba] and assumed it was both written and pronounced with a [d]/"d". As a teenager, I looked it up in a dictionary and found out it was actually "stara baba" ("stara" meaning "old" in many Slavic languages). As a child he had interpreted the [r] as [d] or 'tap' and was steadfast in that belief every time he encountered it afterwards.
So then how do we get the ending of the word "frederika" from "terroristas"? Again, quite easily, and probably listener error again. The two "s"'s fall out - "istas" -> "i(h)ta(h)". We are left with the mutation (or listener error) of [t]->[k]. This is also an incredibly common kind of mutation and also goes in the opposite directions. (Hint for those learning Italian - a lot of those "-tt-" patterns in words were originally "-ct-" in Latin: look at "perfect" and "perfetto", for example.) So it's not surprising that when the brain can't figure out a completely unfamiliar or foreign phrase that you think you hear one thing when it's actually another. On later listenings, I heard the [t] instead of the [k], but the first time I swore it was "Frederika" because that was a cluster of sounds (aka a word) that made sense to me, even if it didn't match the rest of it. (Incidently, there is a word "kolo" in the Slavic language I'm learning right now. It means "bicycle" and, in compound nouns, "wheel": cf. kolodvor, literally "wheel hall", translated as an old-fashioned term for "railway station". If you go to Croatia, probably you will see Glavni Kolodvor (lit. "head wheel hall"), the central train station in Zagreb.)
So, to wrap things up, there are completely logical linguistic rules that dictate how you may hear and misinterpret the Spanish phrase that is part of the Harlem Shake meme. The real clue to the whole thing is that [s]->[h] is a common element of some varieties of Latin American Spanish. The phrase seems to come from a performance of a poem by Hector Delgado. Delgado comes from Puerto Rico. And one identifying element of Puerto Rican Spanish? You guessed it, [s]->[h].