ideambulate

Language can have “natural flow” to it. Linguistic flow is a tough-to-describe quality that makes speeches interesting, books fascinating, and lectures bearable.

Good linguistic flow highlights the important parts of a message, keeps readers/listeners awake and engaged, and generally improves communication. Bad linguistic flow makes ordinary thoughts tough to understand, puts people to sleep out of boredom or frustration, and generally hinders communication.

We have computerized spell-checkers and grammar-checkers to keep other critical aspects of our communications understandable, so why don’t we have flow-checkers? I think this need is currently unmet because flow appears tough to parse out in an automated fashion because of the highly contextual “intuitive” nature of flowing spoken messages.

I’d like to propose using “cadence” informed by “meta-emphasis markup” as a computing-approachable proxy for flow. By cadence, I mean the patterns of stressed and unstressed syllables and the rhythms of consonance and assonance in a message. By meta-emphasis markup, I mean the use of document mark-up that helps clarify the tone and significance of certain phrases. Just think of them as “notes in the margin” that a speaker writes to remember how to say certain parts of a speech.

So, an ideal flow-checker would work like this.

1. A writer types up a document.
2. The writer adds markup notes to the document in certain places needing emphasis or any other sort of special attention.
3. The flow-checker goes through the document and rearranges phrases and replaces words using a specialized thesaurus so that the cadence of the message matches what the markup notes wanted.
4. The writer reviews the flow-checker’s proposed changes to make certain things still make sense with the improved word choice and order.

Given the same original document with most of the same in-text-body markup notes, one could specify the particular tone, dialect, or accent that the message should be made to flow in. Based on the chosen tone/dialect settings, the flow checker would choose to use (or not use) certain words when it goes to optimize the flow. For individual users, this could also mean using some sort of spoken-script capture methods to determine their personal speaking styles so that the flow editor can make smarter word-choices.

Alternately, a politician’s speech-writer could make a straightforward speech, and then use the flow-editor to make customized versions of the same speech for different audiences. Maybe you want to use more polysyllabic words when talking at a university dinner party, but want to sound familiar and down-homey in your turns of speech when meeting a concerned-citizens group for an informal lunch. With a good flow-editor, rephrasing language could be a whole lot less effort for non-English majors and an accessible option for cross-country speakers.

Hmm, running short on time and long on words. More on the meta-emphasis markup possibilities later.