Monday, May 30, 2011

The jonquil glade mystery / how do you know what you know?

Sometimes things make you immediately think, “whodunit?” 

Today as I walked through a nearby open space preserve  I found a glade full of jonquils that were blooming with wild abandon.  They must have been planted some time ago—jonquils aren’t native to this part of the world, and they’re not transported by accident.  They were densely packed in no particular pattern, just the bulbs bumping together in an otherwise quiet grassy spot.  It a very tight cluster of flowers, overwhelmingly transcendent scent as I stood there admiring them.

But they made me wonder: Why are they here?  There’s no way this is an accident, and it’s too far from anything else to be a gardener’s idea of a planting folly, something done just for the unexpected fun of it.  So  I started poking around through the woods by the jonquils, finding a bit of foundation here, an old water pipe there.  Clearly, this was the remnant of an old house that was no longer there. 

What was the story with the missing house? 

Back home I did a little web research.  Looking at the aerial photos of the place on Google Earth I see blank spots on the hilltop.  A little more quick digging told me part of the story.  It turns out that there was once a ranch house there.  It was the “Casa Maximo Martinez"--a sprawling ranch home with six bedrooms, five bathrooms, a walk-in freezer, cavernous hallways, a huge living room with a fireplace, a large dining room with bay windows and a swimming pool. The property was torn down in 1997 after attempts to make it into a youth hostel were thwarted by local residents (including former HP CEO John Young).   They didn’t like the idea of all those young ruffians hanging about in their neck of the woods.

The story goes that in 1833, Gov. Jose Figueroa granted one square league (around 3,500 acres) of foothills known as Rancho Cañada Corte de Madera to partners Domingo Peralta and Maximo Martinez.  After the death of his wife in 1834, Peralta sold his share to Martinez, later enlarging the rancho to around 20,000 acres, including most of what is now Stanford.  (For his part, Peralta moved back to his father’s rancho a few miles south to Rancho San Antonio, near where Interstate 280 intersects Foothill Blvd.)   

The house was apparently built in early 1948 by John Marthens, and aerial photos I found on Google Earth from 1948 show the large house and a nearby oval racetrack that must have been for horse racing or training. 

So the jonquils must be the Marthens’ family garden.  From the looks of it, the patch hadn’t been tended in at least 20 years—that all makes sense. 

And then it occurred to me:  Sometimes there are things that make you think beyond “whodunit?” to “how’d I know that?” 

In this case, I’ve seen what happens to untended bulb gardens.  If you leave them alone for more than a year or two, the bulbs start budding daughter bulblets (also called “offsets” or “bubils”) all around the perimeter.  Even if you start with the original bulbs 6 inches apart, in only 2 or 3 years, the space between them fills in.  Over the course of 20 years, the edge grows outward and makes a ragged perimeter and a pretty solid mat of flowers in the middle.  Of course this probably isn’t what your gardening soul wants (the flowers in the center get compressed and deprived, so their flowers aren’t as strong or large), so gardeners tend to divide and separate. 

But I know this because I’ve seen bulb gardens be left unattended for years at a time.  So it’s by direct experience of daffodils and tulips. 

So how do I know about the history of the Rancho and the story of Domingo Peralta and Maximo Martinez?  The answer’s obvious... I just looked it up. 

But how do I *know* that story is true?   Or, more generally, how do you know that *anything* you’re told (or read) is true / correct / right?

I realize that there’s a huge literature on epistemology, the philosophy of how-you-know.  I’ve read some of that (I started doing my PhD thesis on the epistemology of knowledge representations), but I also find it mostly incredibly obscure, abstruse and abstracted from the question at hand… which is this—operationally speaking, HOW do I know something? 

It seems that it breaks down into direct knowledge (such as my understanding of how gardens of jonquils crowd together over time), and knowledge I’ve read / looked-up / been told… the *indirect* kind of knowing. 

Okay, stay with me now, this gets interesting right around here. 

HOW do I know to trust what I’ve read /looked-up/been told?  (I’m going to abbreviate this as RLBT, shorthand for “indirectly learned knowledge.”)  The answer is that we all develop an intuition about what we RLBT and how much to believe it. 

The problem is that this is largely automatic.  The process of RLBTing and deciding to believe something is so well-practiced that we do this constantly beneath the level of conscious perception.  When I hear something on Fox News, I’m immediately skeptical (without thinking much about why); when I read something in the NYTimes, I’m immediately believing (again, without much thinking about why). 

It makes sense that we operate this way.  One of the huge benefits of language is that we can learn from other’s experiences without having to go through those experiences (possibly dangerous or tedious, or both) ourselves. 

But something fundamental has changed in the past 50 years.  Once upon a time it was fairly difficult to produce something that others would read (and I have a broad definition of “read” in mind, including all forms of publication).  Even something as simple as writing a pamphlet took a bit of gumption, money and time to do.  All of that acted as a kind of “credibility bar” above which writers would have to pass.  Want to publish a book?  You’ll spend a good deal of time (years!) actually writing the book, then you need to convince a publisher to front the money, then distribute the book, etc etc. 

Now, of course, the cost of production for written materials (including video, audio, etc.) is basically zero.  If you want to write a book, you can borrow a library computer, type something up and push it to a web site with almost no cost.  With a little polish, it can seem as valid as anything else out there on the web. 

So how do we evaluate the trust-worthiness of what we find now? 

In some sense, it’s the same as it’s always been—you’ve got to know what to accept based on two factors:  (1) does it come from a source with a positive reputation?  And, (2) does it cohere with what I already know about the world? 

The “positive reputation” part seems simple enough.  Have you heard of the source and do you have reasons to believe that what they’re saying makes sense?  Problem is, with the low cost of publishing, MOST of the sources that might be useful don’t have any reputation for you to evaluate.   For example, when I looked up what the daffodil bulblets are actually called (“offsets”) I found the University of Illinois Extension program website.  Should I trust them?  They’re a university, one I recognize, and I know from previous encounters that they actually have an “extension” program that specializes in this kind of information.  Great!  So I did a quick check on that term by searching for [ bulb offset ] and found that it’s in fact the correct term, LOTS of other sites use “offsets” to describe bulblets as well.  Consensus rules. 

Of course, not all sources have uniformly high credibility in all areas.  I trust my mother to tell me true and interesting facts about our family history, but I wouldn’t trust her explanation of cryptography one tiny bit.  Likewise, I often trust university websites for information about basic science questions (e.g., bulb propagation), but wonder about their non-scientific content, such as opinions about the Iraq war.  Maybe I’d trust their opinions and their data contained within opinion pieces, but only if it “makes sense” to me—that is, if it aligns with what I already know about the world.   (And, as you can imagine, not all university web sites are created equal.  I don’t believe anything I read about evolution from Oral Roberts University— 

“Coherence” is the second big piece of validation.  When we RLBT something,  we all instantly determine if it’s coherent with what we already know.  This evaluation can be shallow (“yeah, sounds the same…”) or deep (“all the points align”), but it happens pretty quickly.  And when it sounds the alarm, the concern sometimes takes some figuring out.  Why does this feel fishy?  What’s not right here?  What doesn’t *cohere* with what I already know? 

The trick is to learn when something’s out of alignment (not-cohering) in a way that’s fundamentally flawed versus something that’s out of alignment because you don’t have accurate pre-existing information.  This is the great role of education—to teach the skills that let you evaluate whether you’re learning something new and valuable, or if you should reject what you’re RLBTing as dissonant. 

For instance, when I looked up the bit about the Rancho Cañada Corte de Madera, I already knew a good deal about Spanish land grants.  As a side-hobby, I read histories of California and know that large grants of land were given to Californios (such as Peralta and Martinez) in the 1830s.   And I knew I could even probably track down the original claim.  Sure enough, even though my 19th century Spanish isn’t all that it could be, I know enough to recognize the landmarks on the map and know that it’s all about this part of the Bay Area.  

So it all fit—it *cohered* with what I already knew.  What’s more, the map is from a place I’ve heard about (Calisphere, run by the University of California’s Bancroft library).   

If you look at the image of the claim, there’s a museum label stuck on the right side of the image.  That’s credibility enhancing, as that’s pretty clearly a museum cataloger’s annotation. 

IF, on the other hand, the compass rose had the letter “W” marking the western direction, I’d be right to be suspicious.  (“West,” in Spanish is “Oest,” hence the letter “O” marking the western direction). 

It all fits.  So, without much thought, with a merest twinkle of my brain, the map, the claim, the name of the region… it all slips into my memory as a fully authorized belief.  I find it credible, and it’s written down in my neocortext. 

On the other hand, on that same time as I was looking at the patch of jonquils on the former Rancho Cañada del Corte de Madera, I also walked over a trail that has a huge number of  clam shell fragments glinting white against the dark brown earth.  This had always puzzled me, as the trail is about 8 or 9 miles from the bay.  What were the clamshell fragments doing there?

I happened to run into two park rangers and asked them, “what about those clam shell fragments up on the trail?  You don’t suppose it’s a shellmound do you?” 

The answer came back a bit too quickly.  “Nope.”   One ranger gave the other a sidelong glance.  “No way there’s a shellmound up there.”  The second ranger agreed, again, a bit too quickly.  “No sir.  There never was a prehistoric village up there.  No way.” 

Uh… Okay.  Their hurry to tell me there wasn’t anything there made me deeply suspicious.  It wasn’t what they said, it was the *way* they’d said it—in a rush, a little dismissively, with a sense of wanting to move onto something, anything, else. 

They’re a credible source, they know a good deal about what’s in the open space preserve, but their way of saying it left a lot to be desired.  I had a few shell fragments in my hand to show them, but thought that discretion might be a good move at this point.  No need to press the issue.

The next day I went back to return the shells to the non-existent shell midden, located on a beautiful overlook above the bay.  It’s kind of a schlep to lug clams up that far, but the view is utterly worth the trek.  And besides, a couple of hundred years later, there would be jonquils at bloom just a bit downhill and to the left. 

Believe me.


  1. wonderful post. Thank you :)

    i have similar insights when i look at websites, only looking at the design i can tell if they are corporate, personal, spammy, blog, etc. I think that comes from experience (i started web surfing and soon went to work as a webdesigner).

    my personal TOP 6 about trust is:
    *repetitiveness (how much i can verify it practically in predictable steps).
    *emancipation factor (how much trusting you ties me to the knowledge only you have).

  2. RRR - Can you say a bit more what you mean by "predictability"? Do you mean "I can predict it based on what I know about this area?"

    Also, for you, what's the difference between "repuation" and "authority"? (I think I know, but I'd like to hear your thoughts on it.)

  3. predictability: yep i don't know if it's translated correctly from italian, mostly is what you said plus the egocentric concept that someone has to "already know" what they are searching for. Sometimes is a positive value (like when you learn something that is in the same direction of your thoughts, and you think "ah! i knew it!" and sometimes can be negative (like when you learn a new amazing thing, when you think "oh! i would never expect it to be this way"). in both cases (a-ha factor and wow factor) there is an impression related to the old knowledge. I'm sure there is something else to dig inside this concept, i hope to wrote it clear enough for the moment!

    the difference between reputation and authority is big: reputation may be measured among peers (would bebetter to define what is a peer), while authority may be a perception derived from how the official sources speak (example: depending of the country of origin, everyone has a different feeling about what's authoritative, some governments use violence and scary sentences to set the rules, that kind of violence + scaryness may be misunderstood for authority when some one is mimicing that mood correctly )

    sorry for the long comment!