Human speech is the vocalized form of human communication. Its two components, speech production and speech perception, link human thoughts to the outside world, including especially the written word but also other communication technologies. As noted in the Ohio State University's
Language Files ,
"Spoken language encodes thought into a physically transmittable form, and writing in turn encodes spoken language into a physically preservable form." Not all human languages have a corresponding written form, but those that do give great insight into reasons for the human success as a species.
Speech and language arguably comprise the greatest communication technologies of all time. But exactly what is speech? How is it different from the audio communications by animals? How did it come about in the first place and how or why does it change over time? Like other adaptations of living things, does it appear and flourish, and then die off? Why are some languages disappearing though others are flourishing? And what exactly is the relationship of speech to language?
Normal human speech is produced from air provided by the lungs which creates sounds in the glottis in the larynx. These sounds are modified by the vocal tract to form different vowels (open vocal tract) and consonants (constricted vocal tract). Among
vowels,
most languages have voiced vowels in which the vocal chords are vibrating, but unvoiced vowels
are reported
for the Native American languages Cheyenne and Totonac.
As for consonants, these can involve a sudden release of air (stops or plosives), rapid vibration (fricatives), or other variants and they may be voiced or unvoiced.
A language is a collection of characteristic sounds called phonemes that are used in oral communication. By contrast, an alphabet is a corresponding set of written symbols for the language. An example of an English phoneme representation is shown above for the sentence "Writing preserves speech."
One or two symbols may represent a single English phoneme. Moreover,
in some languages including English, one letter of the alphabet (e.g., p as in park, phone) may map to different sounds. Other languages where there is a 1:1 correspondence between sounds and letters are generally makes them easier to learn.
As further argued in the book by Laurie Bauer (at right), some languages are more difficult to learn than others because some have more 'rules of usage' or a greater number of sounds, in addition to lacking a 1:1 relationship between letters and sounds. A language is also easier to learn if it has an alphabet, which may be a phonographic system (symbols represent sounds, as in English) or a morphographic system (symbols represent words, as in traditional Chinese morphographs). The 4 Chinese symbols below represent the English phrase "dances with wolves" ("with wolves together dance"),
|
 
|
Most (>75%) languages have a word order preference; Subject-Object-Verb (SOV, e.g., Hopi, Lakota) is the most common type and Subject-Verb-Object (SVO, e.g., English, Polish, Russian) is less common. The definition of all acceptable sentences for a language is known as its grammar (sometimes called syntax).
In SVO languages the subject tends to be the most important word in a sentence and can often stand by itself as a grammatically correct sentence. English examples are:
Bill studied English.
What did Bill study? English.
In SOV languages the verb tends to be the most important word in a sentence and can often stand by itself as a grammatically correct sentence. Lakota examples are:
Tuwe wanlaka he? Who did you see?
Wanwichaunyankapi We saw them.
A grammar is expressed as a series of rules or productions. A sample grammar for the English sentences might begin with such rules as:
Sentence -> subject verb object
Subject -> article noun etc.
To find out whether a sentence is well-formed, a candidate string must first be scanned for tokens (sometimes called lexemes). The English sentence above ('Bill studied English') has 3 tokens: 'Bill', 'studied', and 'English' and discovering these in an arbitrary string of characters is usually called lexical analysis. Typically the tokens are fed one-at-a-time into a parser that uses the language grammar rules to determine whether or not the string is well-formed. Note that in the English example the tokens are easy to delimit - white space separates them, but this is not usually the case, as the Lakota example above illustrates.
Given lists of tokens and grammar rules, two long-standing computer programs -
LEX
and
YACC
- can be used to generate the required lexical analyzer and parser. FLEX and BISON are modern free versions of these two traditional tools.
|
A great book on the interface of human society and language is Harriet Joseph Ottenheimer's The Anthropology of Language (right). Though primarily used as a textbook for calsses in Liguistics, this book is easy to read and emphasizes contemporary issues and the applications of linguistic anthropology such as
an emphasis on methodology, new skills and techniques, an extensive set of visuals, and many examples from different languages.
|
 
|
|
|
