Otto's Notebook and the Feed That Never Earned Your Trust

Three years ago I left my phone in the back of a taxi, and it took me about four minutes to notice. Four minutes is nothing, but I still remember what the panic felt like. It wasn’t annoyance. It was closer to vertigo, my hand going back to my pocket three or four times as if patting it again might change the answer. I got the phone back before the taxi reached the corner. Later that same week I lay in bed scrolling a video feed I didn’t even like, for forty minutes past the point of boredom, and when my husband asked what I was watching, I couldn’t tell him. Both moments get filed under the same vague words, addiction, screen time, willpower, and none of those words explains either one.

The philosophers Andy Clark and David Chalmers gave me better words, in a 1998 paper called “The Extended Mind.” Their argument starts with a strange claim on purpose: if something outside your head does the job a memory or a belief would normally do, and does it reliably enough, it counts as part of your mind, not just a tool your mind uses. Their main example is Otto, a man with Alzheimer’s who carries a notebook everywhere and writes down everything he needs to remember. When Otto wants to go to a museum, he doesn’t recall the address. He looks it up in the notebook, the way you’d simply remember it. Clark and Chalmers say the notebook isn’t standing in for Otto’s memory. It is his memory, moved onto paper. Jason Wu’s Beyond the Sea and Clouds uses the same idea to explain why losing a long-time partner can feel like losing part of how you think, not just losing company. Otto’s notebook is the same mechanism without the grief.

Clark and Chalmers weren’t careless about what gets to count, though. Loosen the idea a little and a phone book or a whole library could be part of somebody’s mind. They set conditions. The resource has to be reliably there when you need it. It has to be easy to get at. You have to trust what it says without double-checking. And, the one I keep coming back to, it has to be something you put there yourself and endorsed when you did, the way Otto wrote each entry by hand, meaning to trust it later. Otto’s notebook passes all of them easily. Passing them is what makes the notebook his mind and not just his tool.

My phone passes the same tests for a big part of my life. The contacts I saved, the photos that are the only record of one particular afternoon, the note I typed at two in the morning because I didn’t trust myself to remember it by breakfast: all of it reliably there, easy to reach, trusted without checking, and put there by me, on purpose, the way Otto filled his notebook. So losing the phone in that taxi wasn’t just losing access to information. By Clark and Chalmers’s own standard, it was much closer to losing a piece of my mind, and that’s the only account I’ve found that explains why the panic felt so out of proportion to a misplaced object. It wasn’t out of proportion at all.

Not everything on the phone passes, though, and the phone never sorts it for me. A photo I took passes. The “Memories” slideshow built from that photo by a process I never asked to run doesn’t. A contact I typed in passes. A notification timed to catch me when I’m weakest doesn’t. Predictive text that finishes my sentence with what it statistically expects, instead of what I was about to write, doesn’t either. The line was never the edge of the device. It runs through the middle of it, feature by feature, separating what I put there from what got put there for me while I wasn’t deciding anything.

The feed is the clearest case of that failure, and it’s what explains the forty minutes in bed. Nothing in it was put there by me, on purpose, at a moment when I chose to trust it. Each video gets picked, in the last instant before it reaches me, by a system trying to keep me watching, a goal that has nothing to do with anything I’ve ever endorsed. Otto wrote his own notebook. I didn’t write my feed, and you didn’t write yours. The extended mind idea doesn’t explain doomscrolling by showing that my mind has spread into an algorithm. It explains it by showing where that spreading fails its own test, and what moves into the empty seat.

The obvious objection is that the algorithm isn’t a stranger. It’s trained on me, on every video I lingered over and every one I skipped, and that starts to sound like a kind of endorsement. But there’s a real difference between what Otto did and what a feed guesses about me. Otto sat down and wrote a sentence he meant, as something he’d trust later. A feed reads how long my eyes stayed on the screen when I was tired or bored or lonely at one in the morning, and counts three extra seconds of not looking away as a vote. I don’t endorse the version of me that exists at one in the morning, which is why a perfectly targeted video feels like being caught, not like being understood. Leaving a trail of behavior isn’t the same thing as writing something down on purpose, and Clark and Chalmers’s last condition is about the second kind.

There’s a harder objection, aimed at the notebook itself, and it makes my whole diagnosis look shakier than I’d like. The philosophers Fred Adams and Kenneth Aizawa argued that being reliably hooked up to something was never enough to make it part of a mind. Going from “this is a reliable, well-integrated resource” to “this is thinking,” they said, is a mistake. Their test asks whether the content means something on its own, the way a pattern of neural firing does, or only because a community of language users agreed what the marks would mean. Otto’s handwriting doesn’t pass that test any better than my feed does. Both are marks whose meaning was assigned from outside. If Adams and Aizawa are right, Otto’s notebook was never his memory. It was always an extremely reliable tool, and the word “mind” was claiming too much from the start.

What falls under that objection, though, is only the big claim, that the notebook or the phone literally is mind. Adams and Aizawa were aiming at that step, from being well connected to being cognition. Nothing about where meaning comes from changes whether a notebook entry was reliably there, easy to reach, trusted without checking, and deliberately put there. Something with all four of those still plays a different role in a life than something chosen by a system that wants my attention, whether or not either one counts as literal thinking. The panic in the taxi and the forty minutes I can’t account for are real either way. I still ask the same four questions about whatever is open on my screen: is it reliably there, is it easy to reach, do I trust it without checking, did I put it there myself? What I don’t know is what to call the thing in my feed that fails every one of those questions and still, most nights, feels like mine.

Editor’s note: remembering something and checking a notebook for it don’t obviously feel the same from the inside. Be Otto for a Minute has you do a small version of both, back to back, so you can feel for yourself whether they actually are.