It's common to think of local knowledge as a smaller version of national knowledge.
We imagine information beginning in individual places and gradually moving upward. Enough observations become a survey, enough surveys become statistics, and eventually somebody looking across the whole field can see what no individual community could see alone.
There is truth in that picture.
There is also something important missing from it.
When information is aggregated, it doesn't merely become larger. It changes resolution.
Suppose we want to know whether a region is becoming more or less prosperous. At one scale we can look at employment, income, business formation, housing costs and other measures that allow one place to be compared with another. Those abstractions are useful precisely because they discard many local particulars. If every town described prosperity in its own terms, comparison would become extraordinarily difficult.
Now move closer.
A business owner may know that customers are postponing repairs they would normally have made immediately. A school may notice that the families leaving the district aren't the same kinds of families who left ten years ago. A volunteer fire department may discover that recruiting isn't difficult because fewer people care, but because fewer residents have jobs that allow them to leave work when a call comes in.
Those observations don't necessarily contradict the regional numbers.
They contain information the regional numbers weren't designed to preserve.
That's the distinction I find useful.
Local knowledge isn't merely national knowledge waiting to be collected, because aggregation can't retain every property of the observations from which it was built. Some information survives compression remarkably well. Other information depends on relationships, sequence, unusual cases or knowledge of what normally happens in one particular place.
The same trade occurs in the other direction. A person living in one town can accumulate extraordinary knowledge about that town while remaining unable to see a pattern occurring simultaneously across fifty other places.
Different resolutions reveal different things.
This is easy to understand with something as ordinary as a map. A highway map can show me how to travel from Albany to Buffalo far better than a detailed street map of one neighborhood. The neighborhood map can show me an alley, driveway or footpath that would be useless clutter on the statewide map.
Neither map is more real.
Each preserves information appropriate to a different question.
Human knowledge becomes harder because we don't always remember that we're using maps.
Someone who has lived in a community for forty years may know which roads usually flood first, which organizations actually cooperate during an emergency and which official procedures are routinely supplemented by informal arrangements. That knowledge can matter enormously when something goes wrong.
It can also be wrong.
The road that always flooded first may have been rebuilt. The family everyone assumes will help may no longer have the capacity. A reputation formed twenty years ago may continue circulating after the person or institution has changed.
Experience provides information.
Time can deepen it, and time can make parts of it stale.
That is why I don't want to give local knowledge a special exemption from correction. Being close to something doesn't guarantee that we understand its causes, and lived experience doesn't make conflicting interpretations simultaneously accurate. People standing in the same town can experience the same change differently and tell incompatible stories about why it happened.
Those disagreements are information too.
The value of local knowledge is not that it arrives already correct. Its value is that it can preserve distinctions that disappear when we zoom out.
Large-scale knowledge has the complementary advantage. It can reveal comparisons that no local observer could