8127/frame/fm

 
ii.
Uncertainty.
By nature, uncertainty is associated with the inference
that results from inductive reasoning. When additional observations are
made of the same phenomenon, our uncertainty about the generalization
changes. If the new information does not falsify our previous hypothesis, it
may increase or decrease our confidence in it. A (preferably quantitative)
statement of this uncertainty is crucial to communicate how new work should
be incorporated with previous observations of the same phenomenon.
According to at least one community of experts in this field, this uncertainty
is properly measured by probability (Lindley, 1991).
Unlike
 
 academic
science, which makes predictions about future events,
a
 
pplied
science is concerned with solving practical everyday problems, often
involving inferences about events that have occurred in the past. Archaeol-
ogists try to infer whole cultures and languages from remnants left behind,
while art conservationists examine works of art in an attempt to establish
provenance. Similarly, in forensic science we make inductive inferences about
criminal events that have occurred in the past. We examine a piece of physical
evidence and want to know if it has anything to tell us about a particular
event of concern. It will not tell us 
 
everything
, because every event has more
than one potential cause, even if some of the potential causes are extremely
unlikely. This is one source of uncertainty in inferential thinking in forensic
science.
In the Sherlock Hemlock tale, it is interesting to note that the additional
information provided at the end of the story by the occupant of the house
caused us to reject our first hypothesis of the birthday party in favor of the
alternate hypothesis, the Twiddlebug dance. Without this information, we
would still be convinced of the birthday party hypothesis.
This awareness that inferences are not absolute truths or immutable facts
is crucial for the forensic scientist. The scientist almost never possesses all of
the information about an event, if for no other reason than the physical
evidence represents only one small effect of the incident. For an inference to
be accepted by either the scientist or a fact finder, each is forced to make some
8127/frame/ch07 Page 164 Friday, July 21, 2000 11:46 AM

---

Association and Reconstruction — Inference of Contact
165
assumptions about the remaining information. The scientist may change her
assumptions based on new information and, when she does, her inference
may also change. Another scientist or fact finder may make different assump-
tions, resulting in yet different inferences. It is invariably the assumptions
about which people disagree, not the facts.