Attaining a Native Accent

In their paper Can Late Learners Attain a Native Accent in a Foreign Language: The Age Factor In Second Language Acquisition, Bongaerts, Planken, and Schils begin with a discussion of the ability of non-native language learners to acquire a native accent. They cite many examples of previous studies which show a lack of ability for non-native learners to acquire a native accent past a certain age. The discussion based on the research of Lenneberg seems to suggest that there exists a critical age for native accent attainment.  However, these studies are not uniform in agreement on the precise critical age or its relationship to other areas of language acquisition—such as grammar, syntax, etc. In each case, the studies did show conclusively that near native accent attainment was most likely to be achieved the earlier the age of arrival (AA) into the country of immersion.

The basis of the experiment by Bongaerts, Planken, and Schils was to test the theory of a critical age. To test this, the authors composed three groups of speakers. The first group consisting of native British English speakers who were determined to have no regional accent. The second group consisted of native Dutch speakers whose second language was British English and had studied English extensively in graduate studies, often lecturing in the language. The third group was comprised of native Dutch speakers whose accent was determined to be definitely non-native. Each group was given four tasks and their pronunciation was recorded. The recordings were judged by four native British English speakers from York.

The results showed that the judges made almost no distinction between Groups 1 and 2, and that, in fact, four of the speakers from Group 2 (native Danes) outscored all speakers in Group 1 (native Britons). The members of Group 3 performed as predicted and were judged to be definitely non-native.  A possible explanation for these results is that the judges were all from York and were able to detect slight regional influences in the members of Group 1.  Also, Group 2 was trained in a form of English pronunciation used by British journalists and broadcasters and considered to be very geographically neutral. Another possibility for the nature of the results is that the fact that Dutch and English are both Germanic languages. Perhaps a study that tested the relationship between the speakers of English and another language, such as French, would show greater differences between actual native speakers and non-native learners because of differences in rhythm and stress.

However the study does suggest that it is not impossible, as Scovel would suggest, for non-native learners to acquire a native accent. The authors do not discredit the advantages of second language acquisition in early childhood as opposed to post-pubescent learning, but they do suggest the critical age be instead regarded as a sensitive age.

Language and Nature: An analysis of Frequency vs. iconicity in explaining grammatical asymmetries by Martin Haspelmath

The argument for iconicity in language has been thoroughly explored by linguists in the past couple of decades. In his paper, Haspelmath argues that many have inappropriately used certain types of iconicity in seeking to explain grammatical asymmetries. The three forms he believes are suspect are iconicity of quantity, iconicity of complexity, and iconicity of cohesion. He does not refute other types of iconicity such as iconicity of sequence and iconicity of repetition.

The argument for iconicity in quantity is that the plurals of words are longer than the singular form because they represent greater quantity. Iconicity of complexity represents the idea that complex thoughts and meanings are expressed by complex forms. And lastly, iconicity of cohesion states that a more cohesive form represents objects or meanings that belong closely together. Haspelmath begins each section by explaining the reasons for advocating iconicity in these categories and cites many of the linguists who have written in support of these ideas.

Haspelmath takes the opposing view by saying that frequency and not iconicity is a more appropriate explanation for these asymmetries. He states that he is dissatisfied with the fact that anything that seems to be non-arbitrary is grouped into the iconicity category by linguists when there are, in fact, other relevant explanations as well. Also, he argues that if iconicity of quantity, complexity, and cohesion are purported to be universal explanations, then they should act as such without breaking down with certain data samples.

Haspelmath states that frequency is a better explanation than iconicity with regards to quantity because the singular forms are much more frequent than the plural. He explains that sign systems follow and economy principle where predictability determines length and frequency is directly correlated with predictability. Also, iconicity of quantity does not extend to explain why words for ‘ten’ are often not longer than words for ‘seven’, etc.

With regards to iconicity of complexity, Haspelmath goes through and shows how each of the less complex forms presented in previous research are also more frequent. He then proceeds to show examples where less complex forms are more marked than their counterparts in areas such as number, gender, causation, case, and person.

In iconicity of cohesion, Haspelmath looks at the frequencies of kinship terms, body parts, and other alienable nouns and their possession forms in English and Spanish to make the point that frequency is a better predictor of these asymmetries.

In conclusion, Haspelmath does not attempt to overstate the role of frequency or completely discredit iconicity in these areas. He states: “I conclude that for most of the core phenomena for which iconicity of quantity, complexity and cohesion have been claimed to be responsible, there are very good reasons to think that they are in fact explained by frequency asymmetries and the economy principle. The final result may look iconic to the linguist in some cases, but iconicity is not the decisive causal factor.”

Linguistic Relativity: What do Hopi and Jedi have in common?

“We see and hear and otherwise experience largely as we do because the language habits of our community predispose certain choices of interpretation." This idea, phrased by Edward Sapir, was the basis for a chapter in Benjamin Whorf's 1956 book Language, Thought, and Reality. Sapir and Whorf became fathers of this idea of linguistic relativity. A definition can be found in Wikipedia under a search for linguistic relativity. The linguistic relativity principle (also known as the Sapir-Whorf Hypothesis) is the idea that the varying cultural concepts and categories inherent in different languages affect the cognitive classification of the experienced world in such a way that speakers of different languages think and behave differently because of it.

In his book, Whorf explores reasons for this interpretation in a series of examples related to fire. Whorf spent many years as a fire inspection and prevention expert for an insurance company and was often lead to determine the cause of fires. He shows in many examples how things being "off" the fire and gas drums that are "empty" influence the perception and interpretation of a situation.

Another interesting part of the chapter is a discussion of the Native American Hopi language and how their language gives them different understandings of time, space, relationships, and other things. One fascinating example is a description of the interaction of Hopi thought. Whorf explains that in English, we tend to form "mental surrogates" of actual objects. Our thoughts then interact with the mental surrogate instead of the actual object. He states: “…this may be so only because we have our own linguistic basis for a theory that formless items like ‘matter’ are things in themselves, malleable only by similar things, by more matter, and hence insulated from the powers of life and thought.”

But this is not so in Hopi... or Jedi for that matter.In contrast, the Hopi believe that positive thoughts can affect a plant positively and negative thoughts, negatively. They also employ “overt” participants in activities whose job is to assist the covert participants with positive thought and energetic emotion in hope of influencing a positive outcome. Thus Hopi thought interacts directly with the actual object--a belief, or interpretation that Whorf argues is a gift their language gives them.

Children's Understanding of Emotion in Speech

Can children detect sarcasm? Can children detect the emotion of a speaker? Morton and Trehub published an article in volume 72 of Child Development that seeks to answer this question. In their article they explore the ability of children to detect emotion in speakers through paralinguistic cues. Paralanguage can be described as non-verbal elements of communication such as pitch, volume, rhythym, gestures and other cues. So the question they asked the question: Do children tend to focus more on the content of utterances, or the underlying paralinguistic aspects of voice sounds? This was tested in three experiments.

           

The first experiment consisted of 120 participants. 100 of them were divided between 4 and 10 years old. The remaining 20 were young adults, mostly college students. They were each given 20 utterances, 10 with happy content and 10 with sad content. However, the happy content used sad paralanguage (lower pitch, attenuated pitch, slower speaking rate) and the sad content used happy paralanguage (higher pitch, greater pitch and loudness variation, and faster speaking rate as compared to the sad paralanguage). The participants were asked to judge if an utterance was happy or sad by saying so or pointing to a sign with a happy face or sad face. The results showed that the 4 year olds were most likely to judge the emotion of an utterance based on content and that adults almost always judged emotion based on paralanguage. The results showed that 10 year olds were more likely than younger children to judge emotion based on paralanguage. There seemed to be a correlation between age and the use of paralanguage in determining emotion.

The second experiment was conducted to decide whether the youngest children (four year olds) were able to judge paralanguage in speech at all. To do this, the experimenters recorded utterances in Italian, so the children would not understand the content, and played them to the children, who were asked to tell if the utterance was happy or sad. The employment of happy vs. sad paralanguage that was described in experiment one was used here as well. The participants were all four-year-olds (8 boys and 12 girls).  The results showed that the children performed better than 80% correct. Some of the children had a tendency to alternate between the two options due to shyness, boredom, etc. When these results were discarded, the children performed better than 90% correct. The results of this second experiment seem to suggest that children are very capable of using paralanguage to determine emotion.

The third experiment attempted to determine whether young children also understood the paralanguage in the utterances from experiment one. To do this they pass-filtered the utterances, which destroys many of the phonetic clues and makes the content largely unintelligible, while preserving the paralinguistic clues of pitch, volume, and speed. 20 four-year-olds, 20 five-year-olds, and 20 six-year-olds were used (half male and half female). The results showed that children as young as four were able to pick up on paralinguistic cues. The six-year-olds did better than the two other groups. It is unclear whether or not this is due to better perception of paralanguage or greater ability to sit through a mundane task.

Summary
What's very interesting here is that young children seem to have the ability to decode emotion in utterances by using paralinguistic cues. However, the youngest children studied, when content and paralanguage conflicted, almost invariable trusted the content over the paralanguage.

Language Change in Subcarpathia

The countries of eastern Europe have seen radical border changes throughout their history. During the last century, following major shifts in political power, individuals and communities have struggled to keep a linguistic and national identity amidst the rule of often oppressive ruling governments. One such area of interest is that of Subcarpathia in the modern-day southwest Ukraine. This region was originally part of Hungary. After The Treaty of Trianon following World War I, this area was given to then Czechoslovakia. Other areas of Hungary were distributed to many neighboring countries, causing Hungary to lose about 70% of the land it previously held and about 65% of its previous inhabitants. The region of Subcarpathia was rejoined to Hungary briefly during World War II and then stripped again at the war's conclusion and handed over to the Ukrainian Soviet Socialists Republic. How has this political juggling affected the Hungarian language of this area's inhabitants?

A walk through the region of Subcarpathia today reveals a merging of two worlds. Street and city names are still often written in the two languages of the region. Many villages in this area still maintain a Hungarian speaking majority while much of the official business and education is conducted in Ukrainian. Because linguistic analysis of this kind was not permitted until after the fall of communism, areas like that of Subcarpathia offer linguists a unique opportunity to study language stratification that has occurred over the course of almost a century.

I recently read one such study conducted in the summer of 1996 and published in volume 19 of Multilingua in 2000. The article is titled The Sociolinguistic Stratification of Hungarian in Subcarpathia. Linguists from the University of Szeged in Hungary went to various regions of Subcarpathia to evaluate the manner in which Hungarian was spoken. Nearly 150 adult minority Hungarians were studied, representing various age groups, education levels, and locales. They were given a questionnaire and asked to fill in the missing word of a sentence. In each question there were only two possible answer choices-- one that was considered to be standard Hungarian based on similar surveys conducted in Hungary and one that was considered to be non-standard. The non-standard forms are considered by those living in Hungary to be associated with lower education levels and rural lifestyle. The group hypothesized that ethnic Hungarians of Subcarpathia would be more likely to choose the non-standard forms.

The results of the study seem to support the experimenters' hypothesis. Hungarian-speakers of Subcarpathia were more likely in almost all instances to choose the non-standard form than Hungarian nationals. However, the results raised an interesting point--that sociolinguistic stratification is by no means a uniform process. While there were some trends among age, locale, education level, etc, language change among individuals was largely unpredictable. There are so many factors that play in to this kind of stratification that effectively measuring all of them would be nearly impossible.

The study raises some interesting questions. How does the political language affect the native language? How does this differ in areas with stronger political assertion? What will the Hungarian language of Subcarpathia look like yet a hundred years from now? Will it still be spoken?