(airy music)
- Scientists often forget that
science is a creative process,
and because it is a creative process,
it has fits and starts.
I think partly we get confused because
we like to think of ourselves as so rational
and having a plan,
and it's also because of the way we write papers,
which is retrospectively.
We tend to craft them so they look very deliberate,
because it's easier to tell a story that way.
But that's not the way science feels on the ground,
and it's not the way that it happens on the ground
in my mind, or in my experience.
It has these periods where you've
gone through the best exercise that you can
to come up with a good plan.
You start to execute your plan,
and something changes.
For instance, you've done an experiment
that you thought were going to be able to interpret,
and you can't.
This can be incredibly frustrating
because you've put in a bunch of different efforts into it.
Maybe it took a bunch of time.
Maybe it took a bunch of thought about how to design it.
Maybe it's a really horrible experiment
that wasn't that long, but just was really awful to do.
Whatever it is, you've done it, you're invested in it.
The data is not interpretable in the framework
with which you set out to do that experiment.
That is biology teaching you something.
It's really important to take
those kinds of failure seriously.
- Failure is
maybe even a critical element of success.
The freedom to really succeed big
requires that we also have the freedom to fail.
In baseball, it's okay if you fail
at the plate seven times out of 10.
You can be a superstar.
I think in science we need
to accommodate that sort of attitude.
One of the things that I've thought about
in the science policy work that I do
is looking for ways so that we can begin to,
if not celebrate failure.
That's a little hard to do.
But at least understand the essential part that it plays
in being able to do something big and important.
- I think it's interesting to pay attention to
what people are saying when they say
failure is an important part of science.
It's a necessary part of science.
Usually, they're talking about
going through a period of failure,
but not residing there.
Failure is a transition state
that gets you to where you want to go.
What I see is that, unfortunately, at times,
you see students who will get used to going to the lab,
trying something out, having it not work,
and then going home at the end of the day
and having a sense that that's what science is,
is that process of failure.
That actually doesn't take you anywhere.
It doesn't take us closer to
that apprehension of the truth and the reality
that is the fundamental goal, I think, of science.
In that sense, failure is, in my mental imagery, a verb.
You pass through failure, right?
By doing the failing, you get somewhere else,
and you learn through the process of extricating yourself
from that position of failure.
- A lot of times, we have periods where things are flowing.
Then there are periods where things get stuck.
I just want to highlight that.
Science, in my experience,
cannot be done without getting stuck,
and I would even say, there's a kind of stuckness
that it says, "Finally, we're stuck,"
because that time is a time where
something about our basic assumptions isn't working.
That is a prerequisite to discover something truly new.
Sometimes I say, if A is the question
and B is the answer, then research is a straight path.
The problem with this schema is
if the experiment doesn't work,
it's perceived as something wrong with the universe
and creates this extra stress, not only the frustration.
I teach my students a different schema
that I think is more realistic.
I say, A is the question, B is the answer, and we get going.
And then experiments don't work.
The path twists, turns,
right, left,
until we got a place linked with negative emotions,
and here comes the word, we call it the cloud.
It's a cloud of research.
You can be stuck in the cloud
for a distribution of time that has a long tail.
It could be a day, a week, a month, a year.
You die.
You're reincarnated as a scientist,
you're still in the cloud.
But if you have enough support, then sometimes,
through the mist of the cloud,
you can see a new answer, C, let's call it C.
And then you decide to go for it.
And experiments don't work, experiments don't work,
but you get there.
And then you publish a paper, A or C?
The cloud is an inherent part of research,
'cause it extends the boundary
between the known and the unknown.
Because in order to discover something truly new,
at least one of your basic assumptions has to change.
Otherwise, it's not new.
That means, as scientists, we try every day
to bring ourselves to this uncomfortable place, the cloud.
Notice that I put B in the land of the known,
because we knew about it when we started.
It's our grand proposal, our Ph.D proposal.
It's very important to have a good solid B
'cause it's our direction of where to start.
But if we keep things wide enough,
when we get to the cloud,
we stand a larger chance of finding C,
which is almost always more interesting,
fruitful, amazing, opening than B.
- Many of us who have been doing science for a long time
can think back to a time when we had a terrible setback
scientifically, and it led to some peripheral discovery,
a realization, we really weren't moving toward, initially.
It was the setback that sparked a realization
that was extremely helpful in moving the science forward.
When I was, this was probably in my third year
of assistant professorship and I had my first post-doc.
She was doing an experiment.
She was supposed to be measuring how
synaptic transmission was reduced,
blocked by different concentrations of cadmium,
because cadmium is known to block synaptic transmission.
This is a very well-established phenomenon,
and we were just interested in getting
the quantitative measurements of
how much different concentrations of a cadmium
were blocking synaptic transmission.
She started doing the experiment,
and the cadmium wasn't blocking the synaptic transmission.
This is, you know, violation.
It's like holding a ball in the air
and letting go of it and it doesn't drop.
It's just something that's not supposed to happen.
I was bemoaning that fact and I thought,
"My gosh, my post-doc doesn't even know
"how to profuse cadmium into the bath."
After a while, I started to believe her,
that, actually, the cadmium wasn't having that effect.
Then I started looking at the data with an open mind,
and it was in looking at those results
that we started realizing that there was something
very special about these synapses,
about the way that the neurotransmitter was released
that made it apparently less sensitive
to the cadmium than we would've predicted.
That was an interesting example of
my own expectations being so entrenched
that when a result came in, I saw it as a failure.
I saw it as a mistake.
We're both harboring this notion
that it was experimental error
that was giving rise to this result
because it so violated our expectations.
For both of us, I shouldn't set this up
as though we were at odds with one another.
We were actually on the same side throughout.
It was only when we came to accept this as a possibility
that we were able to see beyond our expectations
into a new set of ideas.
- Taking those challenges in your project
and thinking about them appropriately
I think is part of becoming a real pro in science.
I would say that if I had to describe
what you're learning in doing your Ph.D,
you're really learning two very basic things:
how to engage with a problem, like a big problem,
and break it down into an answerable question,
and a strategy for answering that question,
and the second one is this ability
to become very comfortable with ambiguity.
And it's hard.
Being comfortable with ambiguity is not easy.
Certainty is a much more comfortable place.
But scientists, by definition,
have to have this calibrated worldview
where you're able to be like,
"Well, we have some little ideas about how that might work.
"They haven't been very well-tested.
"I think it's true for now,
"or true to the best of the extent
"that I can think about it."
Or even like, "I don't know how that thing works at all.
"All I have is this weird set of data that tells me that
"we don't know how to think about it."
Being confident enough in your own intellect
and work ethic to say, like,
"I've thought about this as hard as one can,
"and it is still ambiguous."
it takes real confidence.
I think it's a real asset
that scientists can bring to the table,
that kind of humility in the face
of really complicated problems.