The prior state#

Where Week 8 left us#

Week 8’s surprise question exposed a limit: the final answer cannot show what existed before the probe. This week adds a prior state to the ladder record. Pre-commitment means recording a prediction and reason before checking, then preserving them beside the revision. Our new fictional case changes the setting, while continuing the same question about evidence.

How to use this lesson#

Read at your own pace; no presenter is needed. Allow around 30 minutes. Have somewhere to write a few short answers, and keep what you write — this week’s point depends on comparing an early answer with a later one.

You need no gardening background. All situations are fictional.

The situation: a plant in trouble#

Marlow’s tomato plant has yellowing leaves that curl inward at the edges. Before Marlow looks anything up, they write down what they think is wrong and how confident they are.

Marlow’s note: “Probably underwatering — the soil looks dry near the top and the leaves are curling like they’re trying to hold onto moisture. Confidence: medium.”

Only after writing that down does Marlow open a gardening reference.

Enough background to follow the case#

Yellowing, curling tomato leaves can have more than one common cause: too little water, too much water (which drowns the roots and can cause rot), or a nutrient shortfall. Each has a different fix, so distinguishing them matters — but for this lesson, the exact horticulture is less important than what Marlow did before finding out which one it was.

This lesson does not require you to become a gardener. The plant is the vehicle for a question about evidence, not a course in plant care.

The reference Marlow reads#

The reference explains: overwatered tomato leaves often curl and yellow in a pattern that looks similar to underwatering at first glance, but the soil stays wet rather than dry, and the lower stem may show soft, dark patches from root rot.

Marlow checks the soil again. It is wet two centimetres down, not dry. There are soft dark patches near the base of the stem.

Marlow revises#

Marlow’s revised note: “Root rot from overwatering, not underwatering. The soil is wet, not dry — I misjudged that from the surface, which had dried out faster than the soil below it. The stem patches confirm root damage.”

This explanation is more accurate than the first. But notice what we now have: two notes, written at two different times, for two different reasons.

Name the idea: a prior state#

A prior state is a claim — a prediction, a diagnosis, an explanation — that someone commits to and records before they consult a source, get feedback, or otherwise learn the answer. Marlow’s first note is a prior state. The revised note is not; it was written with the answer already partly visible.

Recording a prior state does one specific thing: it fixes what someone thought at that point, before anything else could change it.

Why bother recording it at all?#

You might ask Marlow afterward: “Did you already suspect root rot?” Marlow might sincerely answer “sort of, yes.” Should we believe that answer?

This is not a question about whether Marlow is being honest. It is a question about whether memory of a past belief is reliable once you know how things turned out. There is a body of research that says, often, it is not.

A study on this exact problem#

Psychologist Baruch Fischhoff ran an experiment where two groups read the same 150-word description of a real but unfamiliar historical event, with four possible outcomes listed. One group (“Before”) estimated how likely each outcome was, not knowing which had actually happened. A second group (“After”) was told which outcome had occurred, then asked to estimate — as if they didn’t know — how likely each outcome had seemed beforehand.

Fischhoff (1975), “Hindsight ≠ Foresight,” Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299. Full citation and link at the end.

What the study found#

Across 24 comparisons (different events, different outcomes, different subject groups), the After group’s estimate of the reported outcome’s likelihood was higher than the Before group’s estimate — every single time (p < .001). The average increase was 10.8 percentage points, and over 70% of individual After subjects rated the outcome as more likely than the average Before subject did.

Simply being told what happened made people believe, in hindsight, that it had looked more likely all along — even though nothing about the original passage had changed.

The part that matters most for us#

Fischhoff also found that people were largely unaware this had happened to them. In follow-up experiments, subjects who were told an outcome and then asked to reconstruct their own prior judgment, or to guess what someone else without outcome knowledge would say, still overestimated it — as if their own hindsight-shifted belief felt like it had been there all along.

This is why “did you already suspect it?”, asked after the answer is known, is weak evidence. The person answering is not lying; their memory of the earlier belief has been quietly revised by knowing the outcome.

Back to Marlow#

Marlow’s case is fictional and much lower-stakes than Fischhoff’s historical judgments, but the same mechanism threatens it. If Marlow had not written the first note down, and we asked afterward “did you think it might be overwatering?”, Marlow’s answer — even a sincere one — would carry exactly the risk the study describes.

Because Marlow wrote the note down before reading the reference, we have something the recalled version could never give us: an unedited record of the reasoning available at that point.

Two notes, side by side#

What the comparison shows#

The final answer is correct and better-supported. But if that were all we had, we would have no way to know whether Marlow reasoned toward it, was told it outright, or copied it. The prior note supplies exactly that missing piece: independent reasoning from visible cues, wrong about the cause but not unreasonable given what Marlow could see at the time.

A wrong first answer, honestly recorded, is still evidence — of what someone could infer from limited information, at a specific moment, before help arrived.

A reference for the basil case#

Wilting that appears only in the hottest part of the day and largely recovers by evening is often heat stress: the plant is losing water through its leaves faster than its roots can replace it, even when the soil has enough moisture. Wilting that persists into the next morning more often points to a genuine water shortfall or root damage.

Check your prior note. Then check the basil each morning for two days (or imagine doing so, if you are completing this without a real plant) and see whether it has recovered.

Worked response: revising the basil diagnosis#

Suppose the basil looks wilted at 3pm but has fully recovered by the next morning, two days running. That pattern matches heat stress rather than underwatering: a genuinely dry plant would not fully recover overnight without watering.

If your prior note guessed “underwatering,” that guess was reasonable — wilting looks similar either way — but the overnight-recovery pattern is the evidence that distinguishes them. If your prior note guessed “heat stress,” you had that reasoning available before checking; say so plainly rather than downplaying it.

What your two basil notes demonstrate#

Whichever way your prior guess landed, you now have a recorded prediction and a recorded revision, written at genuinely different times with genuinely different information available. That pairing is the useful record — not just the final, correct-sounding answer.

This single case does not establish that you reliably reason well under time pressure or with real plants in front of you. One instance is a starting point, not a pattern.

A trap to watch for#

Once you have read the worked response, it will start to feel like the heat-stress explanation was obvious. Resist rewriting your memory of your own prior note. If your prior note said “underwatering,” that is what you thought at the time — leave it as written, exactly as Fischhoff’s subjects could not stop their own memories from doing invisibly.

This is not about admitting fault. It is about keeping the record honest enough to be useful later.

Calibration needs a prior claim to exist at all#

Calibration means how well someone’s confidence matches their actual accuracy — for example, whether the things someone calls “high confidence” turn out right more often than the things they call “low confidence.” You cannot measure this from a single case, and you cannot measure it at all from an answer given after the outcome is known, because by then confidence and accuracy have collapsed into the same moment.

A recorded prior claim, made with a stated confidence level, is the minimum unit calibration would need — repeated many times — to say anything at all.

Keep a short evidence record#

Write four short entries about the basil case:

1. Prior claim: what you predicted, and your stated confidence.
2. What you consulted, and when: before or after the prior claim was written.
3. Revised claim: what you concluded after checking, and what specific evidence changed your mind.
4. Limit: what this one instance does, and does not, establish about your reasoning generally.

This record format follows the same observation / interpretation / limit habit introduced in Week 1, applied here to a claim made before evidence rather than a claim made after it.

What this does not settle#

A recorded prior state is real progress: it rules out the specific distortion Fischhoff’s subjects showed, because the claim exists in writing from before the answer was known. But it does not tell us how the revision happened. Did Marlow reason it out alone from the reference? Did a friend point out the stem patches? Did an AI system suggest root rot directly, and Marlow simply confirmed it?

The prior-state record separates “before” from “after.” It cannot yet tell us whose reasoning produced the “after.”

What comes next#

That question — whose reasoning produced the revised, final answer, when help was genuinely available during the task — is what next week takes on directly, using a collaborator: another person, and then an AI system, working alongside the person being examined rather than only supplying a reference to check afterward.

What to take from Week 9#

A correct final answer erases its own path. If you want to know what someone could work out before help arrived, you have to capture that moment in writing, before the answer exists to distort it — because once it does, memory of “what I thought” is not reliable evidence, however honestly it is reported.

Record the guess, the confidence, and the reasoning, before you check. Then keep both versions.

Read further: hindsight and judgment under uncertainty#

Fischhoff, B. (1975). “Hindsight ≠ Foresight: The Effect of Outcome Knowledge on Judgment Under Uncertainty.” Journal of Experimental Psychology: Human Perception and Performance, 1(3), 288–299. Ask: why were Fischhoff’s subjects unaware that outcome knowledge had changed their own judgment? Read the paper.

This reading deepens the lesson; you can complete the activities without it.

Sources and your next step#

Empirical grounding: Fischhoff (1975), above — an original research report, not a secondary summary.

Marlow, the tomato plant, and the basil are fictional teaching examples, not reported observations of a real gardener.

For the complete written activity and a record to keep, open the Week 9 investigation. The lecture notes offer a continuous reading version.