Babble mode in voice dictation


Voice input has dramatically increased my writing output as well as the quality of my ideas. But I think that it’s not as easy as turning on voice dictation and writing an essay from start to finish. There is a particular way or method of doing voice input that I have found to be far more effective, which is what I will describe in this post.

The pitfalls of dictation

When you think of voice input, you might think of dictation. It’s a form of transcription: an attempt to accurately piece together the words being said and faithfully reconstruct them into text.

Courts and legal depositions benefit from accurate transcription. And in the past, if you were exceptionally rich and exceptionally busy, you might dictate your letter to a scribe who would jot it down for you. These are the use cases that voice dictation has been designed to replace.

A key feature of this mode is that dictation is a one-way process where the goal is accuracy. Do-overs and backtracking aren’t really possible: the over-arching goal is faithfulness to what was said. This is suited for letters and messages, but I think this is not all that voice input can do.

Consider sketching on a page, where various ideas or points are connected by drawing arrows or shapes or groupings. When sketching, you aren’t quite sure how to structure your thoughts before you start. And sketching is, in some way, the act of figuring this out. There is something highly nonlinear about this process where your thoughts backtrack to previous ideas and test the strength of old conclusions.

Good conversations can be something like this, with both participants in more of a ‘dance’ than a single-pointed journey towards a predetermined conclusion. And I think that voice dictation can, with some effort and learning, become something like this too.

An alternate use of voice input: babble

The primary thing to know about speaking is that we can speak very fast, and in comparison we type rather slow. This might seem like no big deal: it’s always tempting to think that a sped-up version of a process is a qualitative, not quantitative, difference. But, as it has been said, quantity has a quality all of its own. Fast input can yield surprisingly large rewards.

Speaking allows words to be put down cheaply, almost at the speed of thought. In that way, voice dictation allows written text to become even closer to pure “thought-stuff”, something to be created, manipulated, and discarded at will. You can speak an entire paragraph that can then be deleted with no sense of loss. This is fundamentally less true for typed text, where each sentence is more substantially a time investment.

Babble is a way to use voice input which takes this crucial difference into account. When I babble, I open a text document, start voice dictation, and then I get up from my chair and start speaking at length, for a minute or two. You’ll be surprised with what you can say in two minutes, and how long the time period can seem in the moment. The longer you can speak without running out of ideas, the better.

Note that there is no penalty for bad ideas, since anything can be deleted. In this way, babble is fundamentally a mode of exploration. The aim of babbling is to generate material to be edited later on. Speaking without planning means you will inevitably make mistakes, but ignore the mistakes when you do make them: you can just say it again instead of leaving babble mode to start editing.

Typing is slow, but babble is fast. Whereas the time commitment of typing means that sentences have to be considered well in advance, babble allows you to circumvent that entirely. There’s no sense being frozen at the computer trying to figure out how to phrase something: you can just say every permutation of the thought. At least one of these is the correct one, and it’s easy to delete the rest.

Babbling means being seldom blocked by the blank page. Even if you have no good ideas, you just say the bad ideas instead. I have found that to be an effective way to get past the intimidating quality of a blank page.

When you come out of babble mode, you might be dismayed to see pages upon pages of half-garbage. This is not as bad as it seems: babble entails editing and wrangling the text to find the core thread of ideas which you are expressing. But it also entails making sure you have plenty of material to do this editing with. It’s counterproductive to have standards so high that thoughts are rejected before they are even written down into text form, where they could have been more easily manipulated and improved. If the bar is too high, nothing can be said at all. The point of babble is to avoid this, and to set your quality bar far lower, in the personal and private interaction between you and your computer.

There’s always a temptation to edit, but to babble means to lean heavily on speaking first. The right mindset is something like “this needs editing, but not yet”. Some of the most productive sessions are walking around in my room, verbally exploring different descriptions of an idea, trying to coax myself to speak for as long as possible. When I sit back down, I often find that I have more ideas than I thought, and that some of my better ideas were being blocked just because I was inaccurately assessing their quality.

Notes Your computer can probably do voice dictation, and if not, it’s pretty easy to install a program that supports this. Android phones are really good at transcription, which can be started by opening up a new note and pressing the microphone button on the keyboard. On desktop web browsers, I can recommend the voice dictation feature built in to Google Docs.

It costs nothing monetarily, and this way of babbling has made me a fair amount more productive when writing. I would encourage you to not get discouraged if it feels awkward, especially for the first few times. Just keep trying to talk, keep trying to babble. This might take a while to get used to, especially if you are used to being critical of yourself, so I’m eager to hear about your experiences, if you do try.