Now that we're neck-deep in the US Presidential campaign, one can't help see the two nominees giving political speeches every time one glances at a TV -- at home, walking through an airport, in a restaurant or bar, etc. The other day, it struck me for some reason how strange it is that, here in the 21st Century, the speech has survived as such an important form of communication.
The general idea of a speech-as-the-reading-of-a-text has always been odd, but it's especially odd in today's world of Teleprompters. Why would one read a prepared text out loud to an audience, instead of simply distributing the text, itself?
On all dimensions of formal information theory, giving a speech is sub-optimal. Slower data transfer rates. Lossy data transmission. No random access. No archival attributes or functions (so review is difficult, unless reversion to the text is available). And, unless the text, itself, is made available, all non-electronic recordings of the speech (e.g., written notes) will be highly lossy. And, finally, if you are going to record it for later review, why not just distribute the text in the first place (actually, in the political context, the text often is distributed prior to the speech, further mooting -- from an information theory perspective -- any benefit of reading the text off the TelePrompTer).
This was all true back when speeches were given from notes, and certainly true if the speaker simply read the prepared text of a speech. It's even stranger, and in some weird way, more fraudulent, when the speech is read from a TelePrompTer. Such a speech suffers from all the aforementioned defects of a speech that is read, word-for-word, from a text, but attempts to hide that fact (in plain sight) and, instead, give the illusion that the speech is extemporaneous.
Of course, speeches have functions other than transmitting information. Indeed, transmitting information, at least truthfully or impartially, may intentionally not be any part of the purpose of a speech. Speeches, for better and worse, are often used as "motivational" tools, and, in those cases, the information-transmission defects of a speech are not as salient. Indeed, in such cases, they may not be defects at all.
Except for the use of Teleprompters, all of this is also true for lectures (especially in academic settings). Perhaps even worse, since, unlike a political speech, the main function of a lecture (certainly an academic lecture) really is the transmission of information.
Even from an information theory perspective, speeches probably did make sense hundreds of years ago, when writing and printing were expensive. Back then, giving a speech probably was the way to most broadly and cost-effectively transmit information. By conveying information, even poorly, in a one-to-many format, a speech was still more efficient than the only alternative of talking sequentially to one person at a time.
Today, of course, not so much.
So why do we persist in giving lectures and speeches?