My drug of choice is writing––writing, art, reading, inspiration, books, creativity, process, craft, blogging, grammar, linguistics, and did I mention writing?

Friday, April 11, 2014

The Mailbox: Speech to Text Software and Season 2

Are you afraid speech to text software will make writing too easy? Why is Season 2 taking so long?   
[Remember, keep sending in your questions to chris.brecheen@gmail.com with the subject line "W.A.W. Mailbox" and I will answer each Friday.  I will use your first name ONLY unless you tell me explicitly that you'd like me to use your full name or you would prefer to remain anonymous.  My comment policy also may mean one of your comments ends up in the mailbox. Be aware, I will call you out if you're acting like typing is a superpower.]    

Paul asks:

Are you afraid that speech to text software will make it too easy for anyone to write and flood the market with lots of amateurs?  

My reply: 

Nah. And you shouldn't be either. People who like to hear themselves talk think that shit is the promise land for their writing careers, but it's nothing more than a funny way to type. Right now it kind of sucks anyway, but I assume you mean when it gets even better.

Though, come on. Try not to be elitist about it, Paul. "Anyone" already can write (baring a few physical or mental limitations), and we were all amateurs once. The way you worded this question, it kind of sounds like you think you've got some kind of special gift because you can type, and soon the plebs will negate your mad skillz with their newfangled devil tech.

I'm also going to assume you mean better incarnations of the tech than exist today because right now, even the best programs take about thirty minutes to write a paragraph.

When I was young, I used to love the old Battlestar Galactica. The one from 1978 with the crazy feathered hairdos and the gleaming Cylons, who managed to be simultaneously monotone and cheeky. (This was before the reimagined human-looking ones who fucked everything....but I guess never doggy-style since that would reveal their glowy spines--which seems like a real liability if you're trying to be incognito while still fucking everything, but whatever.) The original show was back in the seventies, so they dealt with really deep science fiction themes like hopelessness, military ethics, and abject poverty by having Dirk Benedict crack wise every thirty seconds or so and everyone smile WAY too much for the last group of humans...on the run...being hunted down and eradicated...by killer robots.

Oh and they also had cutesy robot dogs named Muffit.

                                                              There are also some who believe in Xenu.

Anyway, Adama from this show was played by Lorne Greene, and at the end of many of the episodes he would do a sort of recap/words-of-wisdom/captain's log thing on this computer. As he spoke, the words would just FUCKING APPEAR on the screen. It was some seriously off the hook science fiction tech. We would obviously be in massive space fleets before that shit could actually work. Well, I wanted to be a writer since I was ten. KCOP (which was channel 13 down in Los Angeles) still ran BSG reruns into the late 80's (and then was a staple of the early Sci-Fi [Syfy] cable channel in the early 90's). So I've spent much of my formative years dreaming of being a writer who could just talk into the computer and have my words appear on the screen.

There were even days, I thought that might be the thing holding me back. All I needed to put Stephen King's snail pace in the dust was a fucking Lorne Greene speaky computer machine. Man, I almost would have risked genocide at the hands of reDONKulously sparkly robots just for the chance to get my hands on one of those.

The thing is, now I know how useless that machine would actually be in the broader mosaic of writing as an art form. 

Don't get me wrong; I know people for whom this technology could actually make them ABLE to write. They physically can't write for one reason or another, and software like Dragon has absolutely been a game changer for them. The Workaholic (who stayed for a few months at the Hall of Rectitude as part of the Sydney superhero exchange program) had been hit with a Kamikaze Biostorm Trigger when she killed The Demylinator on top of the Sydney Opera House, and she couldn't write for more than a few minutes at a time some days. This software may make it possible for her to keep publishing books about indie game design.

But dictation isn't writing.

First of all, writing is always going to be work. I'm not worried about it being "too easy" for those dictating bastage plebs who are taking shortcuts around the real genius of my typing, because after ten minutes (when the novelty wears off), writers who want to have careers writing are still looking at getting up every fucking day, working their ass off, and glaring at sentences that just aren't quite doing it. Speech to text isn't really a shortcut; it's just typing with your voice. Once people realize that they've traded in arthritic joints for blown vocal chords, and that they can't just spew out a thirty minute rant and then head out to the mailbox to collect their royalty checks, it's probably going to seem a lot less awesome.

Frankly, I like typing better.  Nothing wrecks the flow like having to stop and fix a homophone or a misheard word, and when I'm writing, I almost never have to stop to think of a word or get tripped up on my own idea.

Second, the actual writing might be easier (sort of) but the craft is still fiction, and that's just as hard as ever. You still need a compelling character. You still need stakes. You still need a rising action. You still need a cohesive plot. You still need descriptions. Your boring ass story with no central conflict won't suddenly become interesting because you are able to just say it into your computer. Speech to text software doesn't make me care about a shitty character who doesn't want anything, nor will it make some emo cerebral navel gazing any less like a textual lobotomy. Speech to text software won't add in symbolism or make a setting work with the themes of the story. Speech to text software isn't going to understand the difference in tone between a close third person narrator and a first person narrator looking back fondly on the events of 20 years ago. The only way it "writes for you" is strictly mechanically.

Third, the best you're going to get down is a draft. You might be able to make redline edits, and then revise big chunks and then dictate down your second draft, but at some point you just have to get in there and muck around with the language. You have to go sentence by sentence, (maybe even word by word) and ask yourself if it's the best way to say what you're saying. Speech to text isn't going to help you with that level of revision. That's actually where you see the difference between writers who get published and folks who just enjoy it as a hobby--the revision process. Ten bazillion people have a story drafted that they will polish up "one day." So maybe speech to text will help more people "win" NaNo, but that's about it.

I think most people who believe that this technology is going to make them better writers think that they are going to fire up those bad boys and just tell a story. But the reason we're writers (rather than on the radio) is because of how much of writing is dependent on a skill set that won't be helped (and may even be hindered) by the mechanics in which the word is conveyed to the page. In a craft where your choice between a semi colon or an M dash can subtly change the story you're telling or an ellipsis can convey oceans of subtext, dictation software is a blunt instrument. When you talk you have tone of voice to help your meaning be clear. In writing, you have nothing but word choice. Writing is also much more precise and speaking more redundant. Anyone "writing" with speech to text will have to account for the fact that writing with one's voice may actually be harder than silently.

Plus, you end up with sentences like this one: "I juss wanted to let you know so that you weren’t sour prized if you come back for shower tomorrow that my cousin is girlfriend, maybe." (Really.)

Cindy writes:

Season 1 took you less than a year, but Season 2 is kind of puttering along in the middle of its second year. I love your blog, but I actually think one of the best parts is all the weird characters you write about working for you and stuff. Having Grendel working in the cafeteria is hilarious.

My reply: 

Thank you so much, Cindy. And thank you for paying enough attention to know these details. I mean it–it's kind of inspiring to find out that more than four people follow that stuff. Usually I kind of feel like I'm laughing at my own jokes when it comes to the plot arcs and the running gags and the weird characters. Those posts get very few page views and my friends mostly just shake their head. That part is definitely me writing for myself.

I'm trying to get a little better about putting a plot arc post at least once every week or two to move Season 2 along, but here's the real reason it's going slowly:

It's because I'm getting better at writing on a set schedule.

During Season 1, most of my posts for the plot arc happened when I had missed a day or needed to miss a day. If I couldn't get my post up, the Octorians attacked. (Which was of course a huge coincidence and I am in no way saying that those events weren't absolutely real.) I wrote posts like those as a quick way to get myself out of a schedule screw up or a skipped post (real though they may have been). By season 2, I was pretty much writing in the groove, and I almost never missed a post. So since they did so poorly compared to regular posts, I rarely write them when I'm not off schedule.

I also used to do two posts a day sometimes. It was great for page views, but it burned me out. So the days where I would do one "real" post and one plot arc post are behind me and when I have a choice, I usually go with the real post since it will get page views.

I already know where Season 2 is going. We'll get there eventually.


  1. I have Dragon and I briefly used it when I was in grad school so I could read and take voice diction notes about what I was reading. What I discovered was that it was far harder to remember all the voice commands to use the software than it was to just type down the notes I wanted to make. It profoundly affected my flow. I remember thinking that the software would have been enormously useful for one of my profs - partially paralyzed by polio and unable to use 95% of his hands - but for me, a person who types 80+ wpm, it just wasn't worth it.

    Also, I speak differently than I type. If I tried to speak my book into being, it would not be as grammatically correct and it wouldn't flow as well. Again, if I were paralyzed I would have incentive to learn to speak a different way and to memorize those codes you need for Dragon - but it's just not worth it for me as I am right now. I haven't used my Dragon software since early 2010.

    1. Yeah, I had a paragraph where I started to get into the actual neurological differences between speaking and writing and how they would literally affect which "mode" you were in using speech to text, but the article was already getting a little wordy so I let it go.

      I assume better programs than Dragon will come along eventually, but I still don't think they'll do the writing for anyone.

  2. I knew I never should have followed The Demylinator to the top of the Sydney Opera House. It was a dramatically appropriate trap and the rising beat of ominous soundtrack music should have warned me that something terrible was about to happen...

    Writing via dictation is more difficult and time consuming than writing via direct brain splatter transfer from brain to fingertips. When using a program like Dragon you have to imagine you're GMing a table top game for 12 year olds whose only language is NOT YOURS. Speaking slowly, stopping to repeat every fourth word.

    Before the Kamikaze Biostorm Trigger I could write, edit and publish an 8 page article a week. That's nothing compared to what Chris is capable of in sheer word power - but it was a huge amount of productivity for me.

    Now I'm excited when I can a full paragraph done a day. By the end of the week I may have a complete page...possibly even two!

    Dragon certainly hasn't done anything about my poor writing and grammar. Which is why Chris is Writing About Writing (and Occasionally Some Writing) and I'm writing about game design :P

    1. Given what The Demylinator was going to dump into the water supply, you made the right choice. I'm glad you have a way to at least do some writing though!

  3. I'm with Cindy. Bring on more of Season 2.... :)

  4. I desperately want the tech for this to get better before the arthritis in my fingers makes typing excruciating or impossible. That's years off, touch wood, but the idea bothers me (even though I'm not a professional writer and feel little call to be, and even though I rather suspect that I'm better spoken than I am written, when it comes to the things I care most about). C'mon future!

  5. Personally, I have to write with pen and paper; typing shuts down my imagination for some reason. I find Dragon excellent for transcribing my writing. Trying to keep my notebook in place and on page while typing was a hassle that left stacks of notebooks untranscribed.

  6. Back in the Dark Ages, we Suits dictated. I am a little to young to have had the joy of significant dictating to an actual person -- instant feedback; I got to do it a few times and it is wonderful. We carted around these belts and law libraries had the machines so you could use them at the Courthouse! Hurray! Then we got cassette tapes! Machines as big as a Stephen King paperback, but our very own that went everywhere. Then! Mini cassettes! Micro cassettes! And then the PC which led to the elimination of the instant feedback and corrections and mind-reading provided by one's very own Legal Secretary. We had to learn to type! And spell! Who knew there was so much grammar? Dictation software sounded so cool! The Good Old Days, right? But it was and remains like dictating to a pool of word processors who don't know you, your clients, your cases and your quirks. So you had to dictate things like "Period. Paragraph. Should the party of the first part comma notwithstanding..." Since I finally had to learn to use a keyboard, even though I still need to correct a lot because I still can't spell, that is slow and boring. Quote hurry comma close quote Jason screamed period quote that's an AK 47 he's got exclamation point hit the dirt exclamation point close quote. No. I can type a full page in the time it takes to dictate that. I do not find it workable. Dictation software intelligent enough to act like a proper Legal Secretary and just go by my tone would mean we had invented a deity that would shortly take over the world. Until that happens, I will stick to my keyboard.