Natural language

MRI of a person speaking.

A natural language is any language which arises in an unpremeditated fashion.


1   Languages

1.1   English

The old French didn't have the "w" sound, so all words that used to have a "w" got a "gu-" instead. After the Consequent, English got both versions. There's "guardian" and "warden", "guarantee" and "warranty".

"Wyvern" evolved from Latin "Vipera".

Originally, "man" referred to mankind, not males. A male was a "werman" and a female was a "wifman". (The same "wer" used in "werewolf".) Eventually we dropped "wer" from "werman" and "man" became gendered. "wer" ultimately dervices from the same PIE root as the Lint word for man, vir.

2   Form

Natural language have parameters for the word order of adjuncts, complements and specifiers appear. [1]

3   Properties

Hockett's design features of human language:

  1. Vocal-auditory channel
  2. Broadcast transmission and directional reception
  3. Rapid fading (transitoriness)
  4. Interchangeability
  5. Total feedback
  6. Specialization
  7. Semanticity
  8. Arbitrariness
  9. Discreteness
  10. Displacement [human]
  11. Productivity (Generativity) (Expressive variety) [human]
  12. Traditional transmission [human]
  13. Duality of patterning [human]
  14. Prevarication [human]
  15. Reflexiveness [human]
  16. Learnability [human]

(01)-(09) are properties of communication systems of all primates.[1]

(10)-(16) are unique properties of human language.[1]

Note: These design features are not independent.

3.1   Vocal-auditory Channel

A communication system transmits signals over a channel.

Example: gibbon calls (vocal-auditory) Example: paralinguistic phenomena (vocal-auditory) Example: bee-dancing (optic) Example: stickleback courtship (optic) Example: instrumental music (auditory, but not vocal) Example: gesture (optic)

Advantage: Vocal-auditory communication leaves the rest of the body free for activities that can be carried on at the same time.[1]

3.2   Broadcast transmission and directional reception

A communication system has broadcast transmission if signs can be received in all directions.

A communication system has direction reception if receivers can localize the source through the message.

Example: bee-dancing Example: stickleback courtship Example: gibbon calls Example: paralinguistic phenomena Example: instrumental music

Note: Consequence of (01). (Stems from physics of sound)

3.3   Rapid fading (Transitoriness)

A communication system has rapid fading if signs fade after sending.

Example: gibbon calls Example: paralinguistic phenomena Example: instrumental music

Not all systems possess rapid fading:

Example: Animal tracks and spoors Example: Written records

Note: Consequence of (01). (Stems from physics of sound)

3.4   Interchangeability

A communication system has interchangeability if a receiver can reproduce any message it receives.

Example: gibbon calls Example: paralinguistic phenomena

Not all systems possess interchangeability:

Example: Courtship motions of stickleback are different, and neither can act out those appropriate to the other Example: An infant cannot reproduce the message of its mother Example: stickleback courtship

3.5   Total feedback

A communication system has total feedback if a sender receives and comprehends messages it sends.

Example: gibbon calls Example: paralinguistic phenomena Example: instrumental music Example: I hear myself say "the dog" when I say so

Not all systems possess total feedback:

Example:stickleback courtship (The male stickleback does not see the color of his own eye and belly that are crucial in stimulating the female)
Importance:Enables the so-called internalization of communicative behavior that constitutes a major position of thinking.

3.6   Specialization

A communication system has specialization if speech serve no purpose except communication.

Example: gibbon calls Example: paralinguistic phenomena Example: instrumental music Example: A dog, panting with his tongue hanging out, is performing a biologically essential activity, since this is how dogs cool themselves off and maintain the proper body temperate. The panting incidentally produces sound, thereby may inform other dogs as to where he is and how he feels. But this transmission of information is strictly a side effect. Example: Echolocation is not specialization

3.7   Semanticity

A communication system has semanticity if signals have meaning.

Example: bee-dancing Example: gibbon calls Example: paralinguistic phenomena Example: The English word "salt" means salt, not sugar or pepper. Example: Calls of gibbons possess semanticity; the gibbon has a danger call and it does not in principle matter that the meaning of the call is a great deal broader and more language than, say, the cry of "Fire!"

Not every communication system has semanticity:

Example: An overheated dog will pant to dissipate heat, and an observer may understand the panting to mean the dog is hot, but panting does not stand for overheated-ness the same way "overheated" does. Example: stickleback courtship Example: instrumental music

3.8   Arbitrariness

A communication system has arbitrariness if signs have no indexical meaning.

Example: gibbon calls Example: paralinguistic phenomena Example: The word "salt" is not salty or granular Example: "dog" is not "canine" Example: "whale" is a small word for a large object Example: "microorganism" is a large word for a small object

Not all communication is arbitrary:

Example: A picture looks like what it is a picture of. Example: bee-dancing (A bee dances faster if the source of nectar she is reporting is closer and slower if it is farther away.)

Advantage: Does not restrict what can be communicated Disadvantage: Is arbitrary

Note: Consequence of (07).

3.9   Discreteness

A communication system has discreteness if each elementary signaling unit is discrete.

Example: gibbon calls Example: If a speaker produces a syllable that deviates from the normal pronunciation of "pin" in the direction of "bin", he is not producing a third word, but just "pin" or "bin" in a noisy way. The receiver compensates if he can or else fails to understand.

Not all communication is discrete:

Example: paralinguistic phenomena (volume: a speaker may raise his voice along a continuous scale to indicate degree of anger or lower his voice along a continuous scale to indicate degree of confidentiality) Example: bee-dancing

3.10   Displacement

A communication system has displacement if participants can communicate about things independent from space or time.

Example: bee-dancing

Not every communicative system has displacement:

Example: gibbon calls Example: Vervet monkeys can signal the presence of a predatory eagle, but cannot refer to the eagle that attacks a week ago; their communication is limited to what is immediately present.

3.11   Productivity

A communication system has productivity if users can produce and comprehend messages that have never been sent or received.

Note: Productivity implies a system can generate an infinite number of signs. Note: If a system has recursivity, then it has productivity. Note: Productivity is possible if users can arrange symbols in novel ways to produce meaning.

Note: Productivity is not disproved by being unable to produce a particular message

Example: paralinguistic phenomena Example: bee-dancing Example: instrumental music Example: Pictures do not have displacement, but are productive

Not every communicative systems has productivity:

Example: gibbon calls (finite set) Example: stickleback courtship Example: Alarm calls of Vervet monkeys

A non-productive system may be called a "closed" system. A productive system may be called an "open" system.

Whatever can be meant can be said.[1] A given language may not have the vocabulary or syntax to say what I mean in that language, but there are no barriers in principle to supplementing the impoverished language or saying what I mean in a richer one.[1]

3.12   Traditional transmission

A communication system has traditional transmission if users must learn it (and teach it).

Not every communicative system has traditional transmission:

Example: paralinguistic phenomena Example: gibbons calls (globally uniform, so genetics must be responsible) Example: instrumental music Example: bee-dancing Example: stickleback courtship

3.13   Duality of patterning

A communication system has traditional transmission if signs are made of and distinguished by meaningless parts, which can be combined in infinite permutations to make novel signs.

Example: "tack", "cat", and "cat" have distinct meaning but are composed of the same three meaningless sounds in different permutations Example: Morse code

Note: Consequence of (07).

Note: Duality of patterning is probably the last property to be developed, because it is unnecessary until distinct element run out.

Not every communicative systems has traditional transmission:

Example: bee-dancing Example: paralinguistic phenomena Example: gibbon calls

# Refutation

Language is only one of the symbol systems humans use to communicate.

Example: A wedding ring signs marital status [4] Example: A "thumbs-up" gesture signals success, approval, or hope [4] Example: A facial grimace may signal disapproval [4]

[1]: Searle, Speech Acts

4   Limitations

Many of the major problems in philosophy are a result of natural language being imprecise.

This is a known well fact exploited by comedians and poets.

"Do you know why it's easy for a hunter to find a leopard? Because a leopard is always spotted." (equivocation)

Precision requires disambiguating terms.

[1]Travis (1984)

5   Natural language processing

Natural language processing (NLP) is a branch of computer science concerned with the interaction between computers and natural language.

Tokenization (= lexical analysis) splits a longer string into smaller pieces ("tokens"). Larger chunks of text can be tokenized into sentences, sentences can be tokenized into words, and so on.

Normalization refers to a series of tasks meant to put all text on a level playing field: converting all text to the same case, removing punctuation, converting numbers to the word equivalents, and so on.

Normalizing has a number of distinct steps:

6   Further reading

7   Types of Phonological and Morphological Change

7.1   Regular Sound Change

  • Conditioned only by other sounds in the same utterance, never by non-phonological factors (grammar)
  • Phonetic change
    • Phonetic change, but no phonemic or allophonic change; all examples of sound x become sound y, and there are no other source for sound y
  • Unconditioned merger
    • all instances of sound x and sound y become sound z
    • cot/caught, hock/hawk in many dialects of English
    • Unconditioned loss - merges with no sound
  • Conditioned merger
    • sound x becomes identical with sound y, but only in phonological environment z
    • eg latter == ladder, but time != dime
    • Conditioned loss - conditioned merger with zero; knight/night
  • Allophonic split
    • Phoneme x develops a new allophone x in environment y, but there is no other source of x and nothing further happens
    • Secondary split
      • allophonic split followed by part of the environment trigger the split to be destroyed; the result is that the allophones created become seperate phonemes because now they contrast

7.2   Development of Phonological Rules

  • If a regular sound change is conditioned, it can apply in some forms of a word's paradigm and not others, or in some words but not in other obviously related words, because sound change is conditioned _only_ by other sounds
  • If that happens, the result is an alternation between the surviving old sound and the new one, and that alternation can be described by a phonological rule.
  • Once phonological rules come into existence, all sorts of things can happen to them:
    • Extended to new environments
    • Restricted so they don't apply in all old environments
    • Reanalyzed
    • Restricted
    • Lost
      • usually occurs word by word
      • rule ceases to apply to some words, then more, until its restricted to a short list of words that have to be memorized
    • etc

7.3   Pure morphological changes

  • Two types of morphological change are common and simple
  • First, alternations in a paradigm can be levelled, not only the loss of phonological rules, but simply by using one alternating in place of another
    • levelling is just applying a certain (new) rules where two could apply
    • eg - Greeks/Grecians, Timorese/Timorean
  • Second, in languages in which major words (nouns, verbs, adjectives) are divided into arbitrary classes which have different inflections that mean exactly the same thing, _default_ classes can develop.
    • A default class is the class to which a word is assigned if there is no reason to do anything else.
    • Default classes usually developed from inflectional classes of an earlier period that were most productive