For now bypassing issues of intention, I turn to the second part of our definition of language. How can communication be systematic?
One way of describing what is systematic is to introduce the notion of a code. Referring to two definitions from the Merriam-Webster dictionary, a code is (a) a system of signals or symbols for communication and (b) a system of symbols (such as letters or numbers) used to represent assigned and often secret meanings. In order for people to understand one another, they must share knowledge of something. Insofar as people share a code (of communication), we can say that they are using the same 'language'. Codes offer a vehicle by which we can encode and decode messages, going between some form ↔ meaning.
Codes as Digital Information
A very powerful theoretic approach that has many answers for sorting out the essential from the non-essential is Information Theory, which became widely known through Claude Shannon's classic paper "A Mathematical Theory of Communication", published in 1948. Information theory has proven incredibly useful, particularly in engineering applications, computer science, and other fields that deal with streams of information. Below is a definition of information from Wikipedia (03/29/2018):
This definition of information is not essentially "linguistic" in that it doesn't refer to what we typically think of as the stuff of language. This definition, for instance, makes no reference to words, sounds, and the like. What it does provide is a very general framework for understanding what happens in language, namely the sending and receiving of messages.
While going into the mathematical details of Information Theory is beyond the scope of this post, some useful ideas can be discussed by just introducing one not-so-terribly complicated picture. Consider the following diagram, which represents some important components of information theory:
Information is any entity or form that resolves uncertainty or provides the answer to a question of some kind. It is thus related to data and knowledge, as data represents values attributed to parameters, and knowledge signifies understanding of real things or abstract concepts.
This definition of information is not essentially "linguistic" in that it doesn't refer to what we typically think of as the stuff of language. This definition, for instance, makes no reference to words, sounds, and the like. What it does provide is a very general framework for understanding what happens in language, namely the sending and receiving of messages.
While going into the mathematical details of Information Theory is beyond the scope of this post, some useful ideas can be discussed by just introducing one not-so-terribly complicated picture. Consider the following diagram, which represents some important components of information theory:
 |
| Information Theory |
For now, consider the left-most two boxes ("Information Source" and "Transmitter") to be "Person A" and the right-most two boxes ("Receiver" and "Destination") to be "Person B. This is a picture of Person A communicating something to Person B. Hence, the arrows are pointing rightwards.
For our present purposes, the two most components are the signal and the noise (source). The signal conveys some information. A signal can represent anything, including some aspect of natural language (e.g. words). Crucially though, it must be something digital. This means something ultimately reducible into 0's and 1's via some code. Digital text is one example of this. A sequence of letters, e.g. signal may be converted to ASCII character codes to become: 115 105 103 110 97 108. Then, this in turn may be converted into binary (0's and 1's).
Noise is what hinders perception of the signal. Intuitively, the common-usage meaning of 'noise' captures this more technical notion. Talking with a friend on the phone, background noise may make it harder to make out what is being said. In the context of a phone conversation, noise will typically involve extra sound that makes it harder to hear "what my friend is saying" (← the signal). Other media will have other sources noises.
Let's consider another example. Imagine that I hold up a bunch of bananas. I then ask, "how many bananas do I have here?"
 |
| How many bananas do I have here? |
Now, let's assume nobody is being difficult, and everyone gives a straightforward answer.
Person A: Six
Person B: Six
Person C: Six
If this scenario happened via speech, in a classroom for instance, each person's response would sound a little bit different. Physically, we're not identical, so the quality of each individual's voice will differ. In text, we could represent speaker differences using different fonts as:
Person A: Six
Person B: Six
Person C: Six
The actual form (sound wave in speech, image in text) of what is being represented is different from each respondent. But, there is presumably, an underlying sameness of something being communicated. As represented above, the identity of the letters (s-i-x) constitute the information of the signal. The font differences are noise.
Recognizing signals is something we do all the time. One application you, dear netizen, have likely encountered of categorizing messy things into distinct categories are Captchas and other making-sure-you-are-not-a-robot checks.
(Image generated at fakecaptcha.com)
Tasks to verify humans, among other tactics, ask humans to find a signal amidst noise in ways that may be difficult for computers to detect.
(Image generated at fakecaptcha.com)
The Signal called 'Language'
Now, we turn to the question: what sort of codes do we use when we communicate?
Here, we will proceed with the answer that language, using our definition of "what is systematic and intentional in communication", is a system (which may consist of multiple systems) of signs. In fancy-shmancy terms, it is a semiotic ('meaning-making') system. The development of semiotics will be treated in a future post. I've written a bit about semiotics relating to writing and speech in another blog post, but that is not required reading here (👹 not that this is required reading either).
Signs pair forms and meanings. Thus, we interpret a red octagon with the word "stop" written on it as meaning we must come to a complete stop. Here, the form is what we recognize a stop sign to be. Its meaning we just stated. Traffic signs are relatively simple as far as sign systems go; they for example always appear alone (I hope). There is no "grammar" to how they may be combined. Language, on the other hand, is composed of signs that may be combined in all sorts of ways; this is the expressive power of language! Additionally, words are signs that often correspond to more abstract concepts (e.g. "freedom").
One salient fact about linguistic signs ("the signs of language") is that individual signs are more often than not arbitrary. Forms correspond to meanings without any apparent rhyme or reason. Why is it that "dog" in English corresponds to that class of loud fluffy things that Chinese people most certainly do not eat in Yunnan province? Yet Spanish speakers say perro to name that same class of on-the-lawn-pooping mongrels. Describing the pairing of forms and meanings for signs as arbitrary is to say that there is no good answer to the question of 'why' (other than pointing to how people have said these things in the past, but this doesn't explain how they came to say dog or hund (German) or what-have-you).
Thus we can use the forms of signs as information. In sending and receiving signals, we can communicate meanings to one another, provided all parties have requisite knowledge of a particular semiotic system. This involves being able to recognize forms and know what meanings they correspond to.