Monday, March 31, 2008

rhotacism in Romanian

The Romanian language has many examples of rhotacism, the change of a consonant into an “r”. In Romanian, many Latin words with an intervocalic “n” and “l” changed to an “r”. Linguists have determined this type of rhotacism is most common in languages which have an alveolar trill as is the case in Romanian.

Here is a list of Romanian words which have undergone rhotacism:

cer (sky)
soare (sun)
ger (frost)
miere (honey)
biserica (basilica)
care (which)
inger (angel)
fereastra (window)

The word for sky is “cielo” in Spanish and Italian and "caelum" in Latin. Here the “l” of Latin is preserved. In Spanish and Portuguese “sun” is “sol”. The word for “ice” in Spanish is “hielo” and in Portuguese “gelo”. It is from the Latin word "gelu". As a result of semantic shift, “ger” now means “frost” in Romanian. The word “basilica” is used in many languages besides the Romance ones. The word “which” is “cual” in Spanish and “qual” in Portuguese. The Romanian word for “angel”, once written “anger” but now “inger” as a result of a spelling reform, corresponds to the Spanish word “angel”.

The word “fereastra” exemplifies a different kind of rhotacism. The Latin word is “fenestra” which corresponds to the Italian word “finestra”. Here the Latin “n” changed to an “r”.

Romanian has many words which changed due to a sound process known as rhotacism. In the majority of cases, an intervocalic “l” changed to an “r”, but in some cases, it was an intervocalic “n” which changed.

Friday, March 28, 2008

The Great English Vowel Shift

The Great English Vowel Shift refers to a major change in the pronunciation of vowels in the English language. This occurred in England between 1200 and 1600. This phenomenon was first studied extensively by the great Danish linguist Otto Jespersen who created the term.

The pronunciation of English vowels shifted during this period. The two highest long vowels, "i" and "u", became diphthongs and the other vowels increased in tongue height. Other Germanic languages did not undergo the vowel changes which English did. For example, the Dutch word for "late" is "laat"; the Swedish, Norwegian and Danish word for "see" is "se"; the Dutch word for "name" is "naam"; the Swedish, Norwegian and Danish word for "house" is "hus"; the Swedish, Norwegian and Danish word for "ice" is "is"; the German and Dutch word for "thanks" is "dank".

The letters "a", "e" and "i" reflect the change in pronunciation which resulted from the Great English Vowel shift. However, the letters "o" and "u" do not. Let us examine this more closely.

In most languages, the letters "a", "e" and "i" are pronounced as in Italian and Latin. That is, the letter "a" has a pronunciation similar to that of the vowel in "calm"; the letter "e" has a pronunciation similar to that of the vowel in "red" and the letter "i" has a pronunciation similar to that of the vowel in "be". In English, though, the letters "a", "e" and "i" reflect the changes which vowels underwent in the Great English Vowel Shift. The original "a" changed to a higher vowel, the vowel which used to be represented by "e". The original "e" changed to a higher vowel, the vowel which used to be represented by "i". Since "i" was the highest front vowel, it could not change to a higher one. The result was that "i" dipthongized. It became the diphthong which we hear in the word "my". The vowel "o" changed to a higher vowel, the vowel which used to be represented by "u". Since "u" was the highest back vowel, it could not shift to a higher one. Rather, it diphthongized to the diphthong which we hear in the word "mouse".

The pronunciation of the letter "o" did not change. It remained the same. A possible reason it did not change is that the pronunciation of the letter "u" palatalized which prevented diphthongization. Palatalization often occurs with high vowels. The vowel "u" is one such vowel.

The diphthong of words such as "found", "sound" and "ground" does not appear to follow the palatal glide in English. For this reason, the vowel "u" did not diphthongize but simply changed to a sequence of a palatal glide followed by a high back vowel. This is the pronunciation which we hear in words such as "university", "union", "use" and "usual". This palatalization of the high back vowel did not occur in other Germanic languages.

Though the pronunciation of the vowel "u" changed from [uw] to [juw], diphthongization did not occur. No English word has a palatal glide followed by the dipthong [aU]. The development of the palatal glide prevented diphthongization. As a result, the vowel "o" could not shift in pronunciation to that of the higher vowel of words such as "do" and "blue". If this shift had occurred, the resulting pronunciation would have been to similar to that of "u".

If "o" had shifted in pronunciation to [u], it would have been difficult to distinguish it from "u" (pronounced [juw]). We know from vowel dispersion theory that when languages have few vowels, they tend to be vowels that are quite different from one another. For example, the three vowels of Inuktitut are "a", "i" and "u". It would be unlikely that they would be the "a" of cat, the "e" of red and the "a" of "sofa" because these vowels would not be so distinct from one another. Likewise, "o" was not motivated to shift in pronunciation to that of the higher vowel [uw] because if it had done so, it would have been difficult to distinguish it from "u".

The pronunciation of the English letters "a", "e" and "i" reflects the changes of the Great English Vowel Shift. The pronunciation of the letter "u" changed but in a different way. The vowel quality remained the same but a palatal glide was added. The letter "o", however, did not change in pronunciation. This one retained its original pronunciation.

Thursday, March 27, 2008

the truth of sentences with coordinating conjunctions

We can make sentences with the coordinating conjunctions "and", "or" and "but" to determine the relationship between the independent clauses of each sentence and ultimately the entire sentence. An exercise which illlustrates this gives a value of 0 to a clause and sentence which is false and a value of 1 to a clause and sentence which is true.

Let us start with the conjunction "and". Consider the following sentences:

Celine Dion is Canadian and Faith Hill is American. {1,1} = 1
Celine Dion is Canadian and Faith Hill isn't American. {1,0} = 0
Celine Dion isn't Canadian and Faith Hill is American. {0,1} = 0
Celine Dion isn't Canadian and Faith Hill isn't American. {0,0} = 0

With the conjunction "and", we see that both independent clauses must be true for the entire sentence to be true. Now let us look at the conjuction "or".

Celine Dion is Canadian or Faith Hill is American. {1,1} = 0
Celine Dion is Canadian or Faith Hill isn't American. {1,0} = 1
Celine Dion isn't Canadian or Faith Hill is American. {0,1} = 1
Celine Dion isn't Canadian or Faith Hill isn't American. {0,0} = 0

With the conjunction "or, we see that only one of the independent clauses must be true for the entire sentence to be true. Though "or" is a conjunction (it connects independent clauses), we can also call it a disjunction because sentences with "or" are true only when one of the clauses is true. Now let us look at the conjunction "but".

Celine Dion is Canadian but Faith Hill is American. {1,1} = 1
Celine Dion is Canadian but Faith Hill isn't American. {1,0} = 0
Celine Dion isn't Canadian but Faith Hill is American. {0,1} = 0
Celine Dion isn't Canadian but Faith Hill isn't American. {0,0}= 0

The conjunction "but" appears to behave the same as the conjunction "and". However, this is not always the case. Compare these two sentences:

Celine Dion is Canadian and Bryan Adams is Canadian. {1,1} = 1
Celine Dion is Canadian but Bryan Adams is Canadian. {1,1} = 0

Though the two clauses connected by "but" are true, the sentence is false. Why is this? The reason is that the two clauses contain the same verb and same adjective. This makes the sentence false because "but" signals a contrast. The verb phrases of each clause must be different. Here is another example:

Celine Dion isn't American and Bryan Adams isn't American. {1,1} = 1
Celine Dion isn't American but Bryan Adams isn't American. {1,1} = 0

The sentence with "but" is false because both clauses contain the same verb and the same adjective. The verb phrases are the same. Since "but" signals a contrast, the verb "isn't" and the adjective "American" cannot appear in both clauses if the sentence is to be true. The verb phrases need to be different.

The following sentence normally sounds better when it is connected by "and".

Celine Dion isn't American and Bryan Adams isn't Australian. {1,1} = 1
Celine Dion isn't American but Bryan Adams isn't Australian. {1,1} = 0
Celine Dion isn't American but Bryan Adams isn't Australian. {1,1} = 1 in a specific context.

Without a specific prior context, the sentence above is false. However, it can be true if it negates a sentence such as the following: Celine Dion isn't American and Bryan Adams is Australian. This sentence could have been uttered to counter the previous sentence "Celine Dion is American and Bryan Adams is Australian". However, this situation is rather unlikely, so we can argue that in most situations it would be false.

The following sentence is absolutely fine:

Celine Dion isn't American but Faith Hill is American.
but {1, 1} = 1

The rules for "but" need to be refined. Here they are:

{1,1} = 1 if the two clauses contain contrasting information such as the verbs "is"/"isn't" or if they contain different adjectives.

{1,1} = 0 if the two clauses contain the same information such as the verbs "is"/"is" or the same adjectives.

We can state the rules for "and", "or" and "but" like this:


{1,1} = 1
{1,0} = 0
{0,1} = 0
{0,0} = 0


{1,1} = 0
{1,0} = 1
{0,1} = 1
{0,0} = 0


{1, 1 with a structure in each clause that fails to express an overall contrast } = 0
{1, 1 with a structure in each clause that manages to express an overall contrast } = 1
{1,0} = 0
{0,1} = 0
{0,0} = 0

These rules help to express the relationship between independent clauses and their respective coordinating conjunctions. The conjunction "and" needs to connect similar clauses to be true. The conjunction "or" requires that only one of the clauses be true and "but" requires that the two clauses express a clear contrast for the sentence to be true.

Tuesday, March 25, 2008

the English past-tense marker

Many linguists argue that the underlying past tense allomorph is /d/. They base this argument on the claim that it has the widest distribution. However, though it occurs after alveolar plosives, i.e., "wanted" and "needed", and voiced consonants, i.e., "healed", is it not necessarily unconditioned after an alveolar plosive and vowel. One could argue that the plosive is voiced because it follows a voiced vowel.

There is another clear fallacy in the argument. The statement that we should choose /d/ because it has the widest distribution contradicts the notion that an alloform should be contextually determined. The number of alloforms is equivalent to the number of contexts in which they occur. In this sense, no alloform has a wider distribution than any other. Unfortunately, many linguistic analyses involve the principle that if it works, it is fine.

The notion of an underlying form seems to imply that the alloforms were derived from an abstract form. However, the possibility remains that all variations were derived simultaneously in a specific context. This can occur as a result of linguistic variation.

However, if we must choose an underlying form of the English past tense marker, it makes sense to choose /Id/. Historical evidence can be given to support this allomorph. In the movie King Henry V, Kenneth Branagh pronounces past tense verbs such as "remembered" with /Id/. Ancient texts often show spellings such as "remember'd" to indicate that a vowel which is usually pronounced has been deleted.

The Danish language, a related Germanic language, has past tense verbs such as "huskede" (remembered) and "elskede" (loved) in which the vowels are never deleted. Norwegian has "husket" and "elsket" for these respective verbs. We can also argue that if /Id/ is underlying, we apply deletion to derive the other forms. Deletion appears to be more common that insertion with respect to phonological processes.

How do we derive the other past tense allomorphs if /Id/ is underlying? The verb "wanted" requires no phonological process. It is simply want + /Id/. The verb "asked" is "ask" + /Id/. Deletion gives us "ask" + /d/. We then apply a voicing assimilation to make the alveolar plosive voiceless. The verb "spilled" is "spill" + /Id/. Deletion gives us "spill" + /d/. To me this seems more plausible and elegant than choosing /d/ as underlying.

Monday, March 24, 2008


Glides, also known as semi-vowels, are neither vowels nor consonants. Though they share the feature -consonantal, they sometimes pattern more like consonants than like vowels. For example, they are -syllabic. For this reason, phonologists are often not entirely sure how to classify them.

In French, the indefinite article "un" is pronounced with a nasal before a noun that starts with a vowel. For example, "un enfant" ( a child) has a nasal segment whereas "un lit" (a bed) only has a fully nasalized vowel. The phrase "un oiseau" (a bird), however, is pronounced with a nasal before the labiovelar glide of "oiseau". Here the glide patterns like a vowel. In the word "wow", the first glide is more like a consonant. It is syllable-peripheral and easily perceived. However, the word-final glide behaves more like a vowel. It is the second component of a diphthong and is not easily perceived as separate from the first component.

With respect to acoustic turbulence or obstruction through the oral cavity, glides are similar to vowels. Plosives have the most obstruction followed by affricates, fricatives, nasals, liquids, glides and vowels. In this respect, glides are very much like vowels.

However, other examples show they can behave like consonants. For example, in Danish the "d" is usually a fricative between vowels as in "bide" (to bite). However, the word "arbejde" (to work) has an alveolar plosive, not a fricative. This is also the case in "burde" (ought to), in which the "d" is a voiced plosive.

In Argentinian Spanish, nasals assimilate to the place of articulation of a glide but not a vowel. For example, "con hielo" (with ice), has a palatal nasal followed by a palatal glide, but "con anteojos" (with glasses) is simply a dental nasal. This analysis may be complicated by the fact that the palatal glide of Spanish is often articulated more like an approximant or fricative, however. Nevertheless, this nasal place assimilation is typically associated with consonants.

In Danish the word "liv" (life) is often vocalized so that the "v" forms a diphthong with the "i", but the word "livet" consists of a labiodental fricative. The vocalization occurs syllable-finally, a position where the glide is difficult to perceive separately from the vowel, but also like a consonant in the sense that it is syllable peripheral. In the word "livet", the "v" is syllable-initial and does not vocalize. Thus, it appears that glides tend to behave more like consonants in syllable-initial position than in syllable-final.

In Japanese, the palatal glide has a relatively long duration and in careful speech does not diphthongize. This is exemplified by the word "hai" meaning yes. Also, the glide has a mora of its own. The mora can be described as a unit which is between a segment and a syllable. In the sense that the glide has its own mora, it behaves as a consonant.

Glides, also known as semi-vowels, are fascinating sounds which are relatively difficult to classify. Though they are -consonantal, thus similar to vowels, they are also -syllabic which makes them similar to consonants. They tend to immediately precede or follow vowels, a characteristic which makes them similar to vowels. However, in many ways, they behave like consonants. They occur syllable-peripherally, they have their own mora in Japanese and block spirantization of the Danish "d" when it occurs between a glide and a vowel. Thus, they often pattern like consonants.

Tuesday, March 18, 2008

Canadian French

Canadian French is different from European French in several respects. One variety of Canadian French, Acadian, is so different that it can be classified as a separate language. I'll focus on the variety spoken in Quebec, a variety that is often called "Quebecois".

One variety of Quebecois is spoken in Montreal and is called joual. The word apparently comes from "cheval", the French word for horse. It can be described as a blend of rural French and industrial English and is often incomprehensible to speakers of standard French.

In pronunciation, Canadian French or Quebecois differs from European French in several respects. Differences can be noted in the pronunciation of vowels and consonants.

Diphthongization is common in Quebecois. The expression "Il fait beau" (It's beautiful), which refers to weather, is often pronounced with diphthongs in the second and third words. This is remarkably different from the European pronunciation.

The nasal vowel in words such as "en" (in) and "sans" (without) is a lower and more fronted vowel than in European French. In the words "ma" (my) and "ta" (your), a low back vowel is used instead of a central vowel.

Some French-Canadians use a trilled alveolar "r" instead of a uvular fricative or trill. This is especially true of western Quebec and of older speakers. However, it is less common than it was in the past. Also interesting is that some Montrealers use an alveolar approximant in syllable-final position as in "Bonjour" (hello) and they tend to be monolingual French speakers! Thus they use an English "r" but don't speak English. This appears to be a sound development resulting from language contact with English.

The plosives "d" and "t" are affricates when they precede high front vowels. Thus, the word "dix" (ten") has an alveolar affricate as does "tu" (you). Also interesting is that "d" and "t", dental plosives in European French, have an alveolar articulation in Canadian French.

The high front vowels "i" and "u" are lax when they precede voiceless plosives. In the speech of some French-Canadians, though, they are also lax before all consonants, including voiced plosives and fricatives. In the French of these speakers, they are tense only when they occur in open syllables such as "vie" (life). In the words "vite" (quickly) and "doute" (doubt), French Canadians regularly use lax vowels unlike Europeans who use tense ones.

Another feature of Quebecois is vowel lowering which occurs when the "e" comes before an "r". For example, the words "guerre" (war") and "hiver" (winter) sound very similar to "guarre" and "hivar".

Canadian French, also known as Quebecois, is a fascinating variety of French which is primarily spoken in the beautiful province of Quebec. I was very surprised while on my first visit to Quebec to learn that the word "bienvenu(e)" not only means "welcome" as in "Welcome to Quebec" but also "You're welcome", the response to "Thank you". This appears to indicate that Quebecois has been heavily influenced by English.

Monday, March 17, 2008

Is the lateral liquid a continuant?

Many Linguistics textbooks classify the "l", the lateral liquid, as +continuant. This was certainly the case in my introductory textbooks. In reality, though, the situation is not so simple. In certain situations, the lateral liquid functions more as an obstruent such as a plosive. Let's take a look at a few examples.

In Spanish the voiced dental plosive "d" is an interdental fricative between vowels and word-finally. These can be classified as less perceptually salient positions, therefore making them weak. The realization of the "d" as a fricative can be considered the result of a weakening process. Where it is preceded by a lateral, it is realized as a plosive. This is the case in "caldo", the word for broth. In the case of a preceding alveolar trill, however, the "d" is a fricative. This is so in "cerdo" meaning pig. In these cases the lateral clearly functions as -continuant, unlike the trill which is +continuant.

In many varieties of English the voiceless and voiced alveolar plosives are flapped in an intervocalic position in which the preceding syllable is stressed, i.e., "city", "letter", "metal", "medal". Flapping also occurs if the preceding segment is an alveolar approximant as in "party". If the preceding segment is a lateral, flapping does not occur as in "faulty". Again the lateral here functions as -continuant.

In many varieties of British English, the alveolar approximant is deleted or realized as a schwa syllable finally when there is no following vowel. Examples are "here", "father" and "faster". With a syllable-final lateral, vocalization may occur (this is the case in a few dialects), but never deletion. This would appear to be a feature associated with a sound that is +continuant more so than -continuant.

In the speech of many Brazilian Portuguese speakers, the alveolar trill is often deleted syllable- finally when no vowel follows. This is so in "falar" (to speak) and "pensar" (to think). However, the lateral, though it may vocalize, never deletes. This is again a quality associated more with sounds that are +continuant than -continuant.

Consonant clusters also provide good clues as to the nature of the liquids. English allows syllable-initial .tr as in "train" but not .tl. This is the pattern English has with respect to plosives which never occur in syllable-initial sequences. Here the lateral patterns as a -continuant. English allows the syllable initial sequence .sl as in "slow" but not .sr. (The word "Sri Lanka" is a possible exception but note that many pronounce the first segment as a voiceless alveopalatal fricative usually spelt in English as "sh"). The sequence .sr does not occur and neither does a sequence of two fricatives. That the sequence .sl can occur shows that the lateral patterns here the same as a plosive. Fricative plosive sequences can occur such as .st, .sp and .sk. This provides more evidence that the lateral also functions as a -continuant here.

It may be so that the alveolar lateral tends to function as a continuant but plenty of examples can be provided to counter this. The evidence is clear: the lateral often patterns as an obstruent and can thus often be analyzed as a -continuant. Perhaps a better classification of the alveolar lateral is that of +/- continuant.

Friday, March 14, 2008

Swedish dialects

The number of Swedish dialects varies depending on the classification that one uses. The two main dialects are the Svea and Go:ta dialects. (I am using the colon to denote an umlaut) I like to distinguish eight Swedish dialects. They are the dialects of Ska'ne (a' is an "a" with a ring over it), Gotland, Go:taland, Svealand, Dalarna, Va:rmland, Norrland and Finland.

The dialect of Ska'ne is one of the most distinct. It has a uvular trill and a number of diphthongs instead of monophthongs. It lacks the pitch accent present in most Swedish dialects.

Gotland has a dialect which is so different that it is often difficult for other Swedish speakers to understand. It is known for its relatively low intonation and use of diphthongs which do not occur in standard Swedish.

Go:taland, a region which includes Gothenburg, has a strongly trilled "r" and relatively long vowels. The intonation is also distinct.

Svealand, a region which includes Stockholm, has an "r" that is often very softly trilled or pronounced as a fricative or approximant. The intonation is high in comparison to that of most Swedish dialects.

Dalarna has a famous dialect. The "r" tends to be softly trilled, the rhythm is relatively regular and vowels are relatively long.

Va:rmland has a dialect that shares many similarities with the Norwegian spoken in Eastern Norway. The alveopalatal fricative is used instead of the palato-velar fricative of other dialects, the lateral tends to be velarized and the intonation is relatively high.

Norrland is knows for its strongly trilled r, the alveopalatal fricative instead of the palato-velar fricative in many speakers, the deletion of word-final schwa, sentence stress and low intonation.

Finland has a Swedish dialect which uses a strongly trilled r, unaspirated voiceless plosives, a voiceless alveopalatal affricate instead of an alveopalatal fricative, and a single low back vowel instead of a low back vowel and a low front vowel.

Swedish has many dialects, but the eight which I have described here are among the best known. Many Swedes and Swedish-speaking Finns speak dialects that make it possible to determine where they are from.

Wednesday, March 12, 2008

syntactic change in English

Syntactic change has occurred in English. Two examples of this are negation and do-insertion in questions.

Though the use of the double negative is now considered incorrect, it was once quite common in English. For example, Geoffrey Chaucer, often called the father of English literature, used many double negatives in his poetry. In "The Canterbury Tales", he describes the friar in this way: "Ther was no man no wher so vertuous". This means "There was no man nowhere so virtuous". Of course this sentence is now considered ungrammatical and can be changed to: "There wasn't a man anywhere so virtuous".

Another syntactic change occurred in questions. Today inversion occurs with modals and auxiliaries but not main verbs. "Can you swim?" and "Is he coming?" are examples of inversion but with main verbs we insert "do". This gives us questions such as "Do you like fish?" and "Do you dance?" In the past, however, these questions were formed with inversion, not do-insertion. They used to be "Like you fish?" and "Dance you?". This is the pattern in all Germanic languages other than English. For example, Norwegian has "Liker du fisk?" and "Danser du?" for the English questions. German has "Magst du Fisch?" and "Tanzt du?"

It may be that structures such as "Do you like fish?" developed from the insertion of "do" in sentences as an emphatic verb. "I do like fish" is stronger than "I like fish". If this was the case, it then inverted to form questions as is the case with English auxiliaries and modals. In any case, the use of "do" in questions with main verbs and the lack of double negation are syntactic changes which occurred in the English language.

Monday, March 10, 2008

Cockney English

Cockney English is a dialect of English spoken in an area of London. It is the English spoken by Eliza Doolittle in the play "Pygmalion" and the movie "My Fair Lady". Cockney English differs from standard English in many respects.

One of the features for which Cockney English is most famous is h-dropping. The words "hotel" and "happy" are pronounced "otel" and "appy".

The "t" is often pronounced as a glottal stop. This is the case in words such as "water" and "bottle". It is sometimes even flapped intervocalically.

The dental fricatives are not used. The word "three" is pronounced "free". The "th" of "brother" is pronounced as a "v". Word-initially, it is realized as a "d". "This" is pronounced "dis".

The dialect also has different diphthongs. The word "day" sounds similar to "die". The first component of the diphthong is the "a" of "cat". The word "die" is pronounced with a diphthong that consists of a low back vowel and a high front vowel. The diphthong in "coat" sounds similar to the diphthong in "down". The first component is the vowel of "cat". In "down" the first component of the diphthong is a back vowel.

Another feature which Cockney is famous for is the vocalization of syllable-final "l" which is often called dark "l". Thus, "rowed" and "rolled" sound similar.

The Cockney dialect is famous for pronouncing the word "my" the same as "me". Cockney speakers say "me book" for "my book". Though linguists claim that the Cockney dialect is not as widely-spoken as in the past, it is nevertheless a very interesting and famous dialect of England.

Saturday, March 8, 2008

regularity of sound change

The neo-grammarians claimed that sound change had to be regular. While this certainly appears to be true in the majority of cases, it doesn't seem to be true all the time. For example, I produce a schwa in the second syllable of "England", "Finland", "Greenland" and "Iceland", but not in "Thailand".

Why is this? Some people probably do produce a schwa in "Thailand" but I don't. I suspect that certain pronunciations are due to what we're exposed to at an early age. I probably heard the pronunciation with a low front unrounded vowel in the second syllable and thus acquired it.

Also interesting is that I pronounce "Slav" with a low front vowel like the majority of Canadians but have a low back vowel in the word "Yugoslavia". In the word "Yugoslavia" I suspect that Canadians use the low front vowel less than in the word "Slav". Though the word "Slav" is in the word "Yugoslavia" the syllable structure is different. In "Slav" it's followed by a syllable-final consonant whereas in "Yugoslavia" it's in syllable-final position because the syllable structure is Yu-go-sla-vi-a. The low front vowel never occurs word-finally so this be the reason that the syllable-final "a" in "Yugoslavia" is pronounced by many Canadians as a back vowel.

Also, I've noticed that I usually don't pronounced the first "e" in "interesting" but do so when I say it emphatically- in other words, when I say it slowly and in isolation. In such cases, I tend to pronounce it. However, in the word "whitener", I definitely do pronounce the "e" but it's possible that this is really a syllabic nasal which shares the same place of articulation as the preceding plosive.

Unfortunately for the neo-grammarians, it appears that sound change isn't always 100% regular. The reality is that language simply doesn't function exactly the same way as math and physics where the rules are absolute. For this reason, I view language as a blend of art and science and not simply science.

Wednesday, March 5, 2008

more about the Dutch "r"

I forgot to mention in my previous post that the Dutch "r" can also be a velar fricative when it is post-vocalic as in the word "sterk" meaning strong. When it is pronounced in this way, it sounds like the "ch" in the name Bach.

the "r" in Dutch

The Dutch language has a number of pronunciations for the "r". Most languages have a single type of "r". However, according to Ladefoged and Maddieson (1996), about one-fifth of the languages that have an "r" have two or more variants. Dutch is one of these languages.

Most linguists assume that the oldest variant in Dutch is the alveolar trill, a sound that is common in many languages. The uvular "r", produced with the back of the tongue against the uvula, dates back to the end of the 1600's. This "r" is often called the French "r" by those who hear it.

Another "r" sound, mostly used by Northern Dutch speakers and especially at the end of a word, is an "r" which sounds like the Dutch "j" or the "y" in the English word "yes". For example, the Dutch word "daar" meaning there is pronounced dai (similar to "die") by these speakers. Some linguists claim, however, that this pronunciation is not standard Dutch.

An "r" sound which is apparently spreading is one similar to the "r" sound in the English words "red" and "rose". Linguists disagree on its place of articulation. Some say that it's palatal; others say its palatal-velar or pre-velar. Perhaps the place of articulation varies depending on the speaker. What is clear is that this "r" is always post-vocalic- it occurs word-finally and syllable finally if the syllable is not followed by a vowel. Speakers who use the approximant use the alveolar or uvular trill elsewhere.  The one exception is the city of Leiden where speakers use the approximant in all positions.

The sociolinguistic data regarding this sound is very interesting: a) it is most common in the western part of the Netherlands; b) women use it more than men; c) children use it more than adults.

The latter two points may be characteristic of sound change in general. Women tend to be at the forefront of leading a sound change more than men and children more than adults. It appears that this approximate is spreading in the Netherlands. In the Flemish part of Belgium, however, this is not the case. Most Flemings use the alveolar trill in all positions. A few, however, substitute it with the uvular trill. The approximant, though, is not used by Flemish speakers. Unlike Dutch speakers, they have not adopted it.

Tuesday, March 4, 2008

trap-bath split

The trap-bath split refers to a sound change which occurred in the English of southern England. It is believed that this sound change took place in the late 1700's. The word "trap" has the low front unrounded vowel and "bath" has the low back unrounded vowel. In my Canadian English, both "trap" and bath" are pronounced with a front vowel.

Actually, the low front unrounded [ae] vowel in words such as "cat" and "that" initially lengthened and then became a back vowel. In northern England, this sound change never occurred. It may be argued that the low back vowel is less marked than the low front vowel. However, in languages such as Spanish and Italian, the "a" is not a front or back vowel but rather a central vowel.

In which environment did this sound change occur? It occurred before usually tautosyllabic voiceless (the syllable-final vowel of "castle" is an exception) consonants and nasals. "Rather", if it is an example of this sound change, is also an exception because the "th" is in the following syllable and is voiced. However, this rule does not apply in all situations. For example, the front [ae] occurs in "lass", "mass", "gas", "fan", "hand" and "maths". In some cases, the use of the back vowel is optional as in "trans" in "trans-Atlantic" and in "circumstance". Some speakers use the front vowel and others the back.

It does not occur before the velar "k" or the alveopalatals "sh" and "ch". Thus, it appears that this sound change only applied before +anterior consonants.

I made a list of 75 words in which the back vowel occurs in speakers of English dialects which have the trap-bath split. This is true of Received Pronunciation and other dialects of England as well as Australian, New Zealand and South African English. In the USA, the trap-bath split can be heard in the Boston accent, but for the overwhelming majority of American speakers, there is no trap-bath split.

Here is my list of 75 words:

sample, path, aunt, calf, rather
shaft, raft, master, plaster, plant
advantage, advance, answer, giraffe, daft
graft, Alexander, demand, Sandra, slander
chancellor, enhance, trance, stance, ranch
chant, enchant, slant, brass, fasten
ask, branch, cast, past, last
fast, mast, path, pass, glass
grass, task, mask, dance, chance
can't, rather, bath, laugh, France
half, class, blast, after, craft
draft, graph, staff, castle, aunt
fasten, basket, casket, flask, rascal
blast, contrast, nasty, pastor, vast
disaster, grasp, grant, command, transport

I then determined the frequency of occurrence of the back vowel before each consonant. Here are my results (percentages are not exact due to rounding) :

before s: 31 occurrences [42%]
before f: 12 occurrences [16%]
before n: 28 occurrences [37%]
before th: 2 occurrences [3%] (this is the voiceless fricative)
before m: 1 occurrence [1%]
before th: 1 occurrence [1%] (this is the voiced fricative)

What do these results suggest? They suggest that the sound change in which the low front vowel became a back front one occurred most often before an "s" (42%) or an "n" (37%). Together these two consonants account for 79% of the examples in the data.

The "s" and "n" are similar in the sense that they are both alveolars- they share the same place of articulation. Also, they are both produced with the tongue tip, making them coronals, and they are both +anterior sounds. The "n" is not voiceless but we do not usually need to specify voicing in nasals because they are nearly always voiced.

The trap-bath split is a very interesting phenomenon which occurred in the history of the English language. Surprisingly, though, I never studied this when I took linguistics at university. I learned this on my own.

Epenthesis in Brazilian Portuguese

Epenthesis in Brazilian Portuguese

An interesting feature of Brazilian Portuguese is the epenthesis which occurs in consonant clusters. For example, the name Edna is pronounced e-ji-na or e-di-na in those varieties which don't palatalize. "Admirar" (to admire) is a-ji-mi-rar and "advogado" (lawyer) is a-ji-vo-ga-du or a-ji-vo-ga-do if there is no word final vowel raising. The word "psicologia" (the p is pronounced) is pi-si-co-lo-gi-a.

However, in the word "atlantico" (Atlantic) no epenthesis occurs. Why is this? It doesn't seem plausible to treat the "tl" as a syllable initial consonant cluster because it never occurs word initially. The syllable structure appears to be at-lan-chi-cu or at-lan-ti-cu in the varieties which do not palatalize. At first I thought that maybe epenthesis could only occur if the two segments were plosives, but in the case of "advogado" and "psicologia" we have a plosive followed by a fricative. I then thought that perhaps the two consonants had to agree in voicing but no epenthesis occurs in "Brasil" (Brazil) so this obviously could not be the rule.

In comparison to plosives and fricatives, liquids have greater airflow through the oral cavity. In this sense they are rather similar to vowels. In many varieties of Brazilian Portuguese the l is vocalized syllable finally in words such as "sal" (salt), "caldo" (broth) and "Brasil" (Brazil) . The r is often deleted or barely audible when it occurs word finally as in "pensar" (to think) and "mar" (sea). In these cases they function more like vowels than consonants. The l, when it vocalizes, can be treated as +continuant rather than -continuant. The rule is that consonant clusters are repaired with an epenthetic vowel in those cases where the second segment is not a liquid. "Atleta" (athlete) is at-le-ta and "prato" (plate) is pra-tu.

Why does no epenthesis occur when the second segment is a liquid? Perhaps it is because the liquid is +continuant and thus sufficiently similar to the vowel, thereby eliminating the need for epenthesis. Epenthesis must only occur in those environments where the two segments are rather different. Thus the rule for Brazilian Portuguese epenthesis is that the two segments of a consonant cluster are modified by epenthesis if they are two plosives or a plosive followed by a fricative. If the second segment is a liquid, epenthesis is blocked.


Hi everyone!

Welcome to my blog. I hope you'll enjoy reading it. I plan to discuss various topics such as literature, linguistics, music and chess.