What does the study of natural communications imply about artificial communications?

Such artificial communications would be the set of signals you might devise yourself.

First of all, the functional classification of behavior by Ekman and Friesen might suggest important elements to include in an artificial system of signals. For example, movements with precise meanings and movements that regulate the flow of interaction between user and computer would seem the most important elements in an artificial signal system for a human/computer interface.

Emblems

The artificial symbols that might convey specific messages, such as "Print" or "Move a window", are most similar to Ekman and Friesen's 'emblems' category. Here are some things to keep in mind about such emblem-like signals whether natural or artificial.

The 'OK' Emblem (not okay in some cultures) - Meaning differs across cultures for the same movement. The same or similar movements and postures can be acceptable in one culture, but unacceptable in another. For example the OK emblem means "alright" or "fine" in North America, but it is an obscene, negative gesture in some parts of South America, such as Brazil. 

The lesson here is that you do not want to include in your set of signals ambiguous signals that are misinterpreted in different cultures because of their prior use.

Behaviors that are appropriate for use change with culture, age, and sex. For example, the gesture for 'shame on you' (rubbing the index fingers together, perpendicularly and down) implies a superior position for the sender versus the receiver of the message.

The 'Shame on You' Emblem -  Shame on you is appropriate for an adult to give to a child, but you would not see an adult seriously give this sign to another adult without offense.

The 'Finger' Emblem - The finger, a notorius American emblem, was never seen employed by women up until a couple of decades ago. This disparity is an example of sex difference in emblems. The point here is to avoid artificial signals that have age, sex, or cultural proscriptions.

The difficulty of facial actions varies and changes with age. If your system includes artificial facial signals, they should be based on easy to perform actions for all age groups. The table below (derived from an article by Ekman, Roper and Hager) shows some facial actions of varying difficulty. Difficulty should probably also be considered for any action performed as an emblem in an artificial signal system.

The emphasis on measuring sign vehicles is not an important criterion for an artificial system. In an artificial system the signals are known in advance and the goal is to make an effective system, rather than the discovery of signs and signals. 'Whatever works' would be the motto here.

Adaptors

If you are considering a set of artificial signals, you would not want to include in your set of movements with precise meanings those that are similar to adapters emitted spontaneously. Your signal might be confused with the naturally occurring adapter. For example, imagine an interaction where scratching one's nose was the command to erase a disk file. You would not want to use this motion because many people scratch their nose without wanting to erase a file. The set of artificial signals should be distinquishable from the usual repertoire of non-verbal behaviors.

Regulators

Finally, the use of signals similar to the regulator category might be helpful in controlling the pace or flow of communications between computer and operator. The hour glass wait symbol is analogous to a listener response on the part of the computer. Some potential messages from the user similar to a Regulator category might be "Keep on scrolling the text", "Pan faster", or "I'm not looking at the display now".