So this one took a turn.
Maybe it’s best to start from the beginning. Unlike Generative AI―which is notoriously a “black box”―I think my recounting of my own thought process will at least show a human being at work. For whatever that’s worth. Definitely an open, urgent question.
1. Thought for this tip: Editing games or arbitrary boundaries focus the mind far better than vague admonitions to use fewer words or write simply.
2. A trick I’ve used for years is to eliminate “orphans,” a single line at the end of a paragraph or a line of a heading that is just a word or two. Eliminating the orphan requires me to revise the paragraph or heading and make it tighter. I also find orphans aesthetically displeasing. Wait, that doesn’t sound great. TBD on that.
3. Wait, is the word “orphan” still appropriate? Or has it gone the way of “homeless”? Do we say, “people experiencing parentlessness”? Wouldn’t make much sense because it’s not going to change. Are children no longer “orphans” once adopted?
4. I wonder how the majority and dissent differ in Grant’s Pass on describing the homeless. So fraught.
5. Seems like hot-button language is moving in the direction of describing the current observed state of things, kind of like the Copenhagen Interpretation of quantum theory. Search that. What’s the underlying reality? Is the cat in the box dead or alive? (It’s both. Don’t really get that. Is it actually both, or is it that we just don’t know? Actually both? WTF? Like multiple worlds? That’s different from the Copenhagen Interpretation? WTAF?) But our language does seem to be getting fluid and focused just on present condition or suggesting it could change. Or at least some people like it that way. Like that professor my daughter had, the one who described the first female movie director as a “woman-aligned human.” WTF? Should I include that? Should I include my daughter’s awesome response to that? TBD. Also, should I note that Michel Houellebecq’s Elementary Particles is what had me thinking of the quantum theory overlay? Do I say I really liked the book? A lot of masturbation in that novel. Among other things. But also quantum theory. Very thoughtful book, but, yeah, well. TBD.
6. Back to orphans, I’m going to ask Co-Pilot for famous orphans. Alexander Hamilton! Yes, of course. And what was that blues song that Eric Clapton did? “Motherless Child”! But not necessarily an orphan, just motherless. Probably an orphan, though. Maybe put in a link to a video of him playing that. How do I get those thumbnailed links? Good god, more complicated than I thought. Forget that. Also, was he cancelled for vaccine stuff? Not OK to like Eric Clapton anymore?
7. Focus! Orphans. The opening tip is to eliminate the orphans―well, that doesn’t sound great either―anyhow, revise a paragraph or heading so that the final line isn’t just a few words. You’ll figure it out.
8. Kind of like the Underpants Gnomes on South Park. Phase 1: Steal underpants. Phase 2: ??? Phase 3: Profit. Phase 1: Find orphans for elimination. Phase 2: ??? Phase 3: Better paragraphs with fewer words.
9. Sometimes it just helps to have those arbitrary boundaries. A defined task. Poetry without meter and rhyme is like tennis without a net. Robert Frost. Maybe I should write a sonnet about this stuff. Or just have ChatGPT write one. TBD. Pretty incredible that ChatGPT can produce a sonnet on anything and lawyers think there’s something extra-special about what they write. Something untouchable, or at least sufficiently difficult that they’ve got some runway before oblivion. Right.
10. Focus! Just the tip. Over the years I’ve found arbitrary limits incredibly useful to force editing. Eliminate orphans, no more than one or two pages for an intro, etc. Sometimes I get a draft and just tell the authors: go back and cut 1000 words.
11. Again, the best legal writing advice must be simple, clear, and actionable. Rather than say “look to eliminate passive voice” or cut down on legal jargon, if I say “eliminate the orphans” or “get it down to a page,” you just appraise everything and make choices. It’s not that hard to find different, simpler ways to say something. You just need to know where to focus.
12. And lord knows most legal writing can get a ton shorter by taking aim at explanatory parentheticals. God, I hate them so much.
13. Though the intro thing is more substantive. That one is about making choices about what is most important. And respecting the reality of limited attention. What did Garner say? You have 90 seconds? That’s like a page, maybe a bit more or less depending on the reader. Also, I find it aesthetically displeasing when an intro slops just over a page, like one line over. How difficult is it to edit these things to stop that? About as difficult as tying your shoes or not having your shirt tail come through your fly. That’s what it looks like to me.
14. Pretty annoying how some lawyers resist or are just slobs. Are we hanging on their every word? They just lean and lean into their assumption that the reader has an obligation to suffer through it. Or they think their prose is a lot more captivating than it really is. Or the subject matter. Not like we can intersperse mention of masturbation for shock effect like Houellebecq! TBD.
15. Actually, I wonder how well ChatGPT would do with simply reducing the number of words in a passage of legal writing?
And so it came to pass that I took the lamentable passage from United States v. Lopez that I featured here and asked ChatGPT to reduce it by 25 words without changing any meaning or citations.
16. F&^% Though sort of validating that it deleted the last explanatory parenthetical entirely.
17. What’s that song, the one about the guy who competes against the machine until he dies? “Ballad of John Henry”! John Henry was a steel-driving man. Oh, Lord.
18. You know, the thing about Generative AI is that it’s a black box. It can’t even explain what it’s doing. We don’t even understand how it chooses.
19. Well, I guess I could ask it.
Here’s a breakdown of the approach I took to reduce the paragraph by 25 words while maintaining its legal integrity:
**Condensing Phrases**: I looked for opportunities to convey the same ideas in fewer words. For example, “For nearly a century thereafter” was simplified to “For almost a century.”
**Removing Redundancies**: In legal writing, some details can be repetitive or implied. The phrase “Court’s Commerce Clause decisions dealt but rarely” was simplified to “Court’s Commerce Clause decisions rarely dealt.”
**Simplifying Language**: Replacing longer phrases with shorter equivalents can reduce word count without changing the meaning. For instance, “as a limit on state legislation that discriminated against interstate commerce” became “as a limit on state laws discriminating against interstate commerce.”
**Eliminating Unnecessary Details**: Where the meaning can still be understood, some descriptive elements can be trimmed back. For example, reducing “with the extent of Congress’ power, and almost entirely with the Commerce Clause” to “with Congress’ power, focusing on the Clause.”
20. F&^% That’s basically the kind of legal writing advice people typically give and get. Actually, it’s basically Strunk & White.
21. So lawyers can write rambling, un-conversational, jargonized crap, and if you just feed the draft to ChatGPT and ask it to cut it by 1000 words or whatever you will get back revisions that embrace leading practices for efficient and effective writing? The worst writers can do this and easily improve what they drafted? And it takes seconds? And you can ask it why it made the edits it did, and it will give perfectly reasonable answers?
Lyrics to “Motherless Child,” please.
Lord, I’m going to the river, get me a tangled rocking chair
I’m going to the river, get me a tangled rocking chair
And if the blues overtake me, gonna rock away from here