When does data become evidence? How much data does one need to falsify a hypothesis? How does fact become established? How many facts make a convincing case? These questions deal with the issue of what constitutes a standard of proof (see Standards of Proof).
Data is factual information used as a basis for drawing inferences or calculation, e.g., color, size, composition, dimensions, weight, age, and the like. Researchers establish data sets in many ways—scientific and nonscientific. In science rules of observation establish the criteria defining the practices and methodology to be utilized in establishing a usable data set. These rules commonly include a taxonomic system, wherein taxons are referred to as data elements, for description of the data to be collected. They also may involve guidelines for the observer regarding indications of "patterning" (occurrences of data elements appearing together too numerously to be random).
Upon establishing a data set then the investigator can proceed to statistically describe the set. Ungrouped data may be grouped by data elements into frequency diagrams. A ceramic data set containing a data element describing the current nature of a ceramic artifact as rims, fragments, handles, and the like could be grouped by frequency of occurrence. This kind of inquiry, which also includes measures of central tendency (mean, median, and mode) and measures of relative standing (percentile, quartile, decile), are primary observations. These are all merely depictions of the data set in terms of its static state.
The next step in examining the data set, to determine the relationships that exist between data elements, is its statistical treatment to obtain measures of variability. This level of analysis discloses what Lewis R. Binford calls "first derivative patterning" consisting of measures of linear (coefficients of correlation) and multilinear (multivariate analysis) relationships. This is significant in that it is the initial procedure for exploring the dynamics of how the world works and mandatory for any scientific explanation. The procedure allows a scientist to formulate principles that may lead to an explanation of the universalities of human behavior. Subsequent steps involve nonlinear analysis. Determinations in first derivative patterning requires an acceptable, statistical certainty as a standard of proof for a proffered fact to constitute evidence. Hermeneutic approaches do not call for this rigor.
A scientist cannot falsify a hypothesis without a showing of acceptable, statistical certainty. This remains at the heart of the new archaeology wherein lies the effort to establish a level of methodological precision that yields findings of the same certainty as in the natural sciences. However, in hermeneutics even a "probable" fact has evidentiary value because every possibility has relevance for the analysis. All relevant evidence necessary to give a complete picture has probative value in either scientific or hermeneutic analysis. An argument is an assertion (a claim of fact, a proposition, or a thesis statement) supported by relevant evidence.
Relevant evidence means all evidence having any tendency to make the existence of any fact of consequence to the understanding of the archaeological record more or less probable than it would be without the evidence. Does a proffered fact possess sufficient probative value to justify receiving it as evidence? Probative value depends not only upon satisfying the basic requirement of relevancy as described above but also upon the existence of some matter of fact. The initial problem has to do with fact. Relevancy, not an inherent characteristic of any item of evidence, exists only as a relation between an item of evidence and a matter properly provable in an investigation. The matter turns on:
How then does a scientist use relevant evidence? The scientific method requires scientists to marshal sufficient relevant evidence (an adequate sample) to attain a high standard of proof. The evidence confirms that what is asserted is reasonably true in fact. If the evidence is not relevant, then it has no bearing on the veracity of the assertion.
This does not necessarily become an issue in hermeneutic analysis. Hermeneutic analysis often employs referral arguments. What distinguishes a referral argument from other arguments lies in the proffered supporting evidence being an established, or trustworthy, frame of reference (an "authority" or some standard accorded significant credibility). The inference in a referral argument consists of an assertion being trustworthy because of its referral to and confirmation by a trustworthy frame of reference.
Of significance to archaeology remains the fact that underlying referral arguments exist in all assertions made from research data and about the nature of the data. This introduces a subtle form of uncertainty into research. When archaeologists find an artifact, or make an observation, they often make an assertion based thereon without always critically examining underlying evidentiary assumptions, but simply refer to frames of reference which may be their own experience or intuition. The implication arises that, absent a careful evidentiary analysis, error may be compounded upon error giving rise to questionable and at times bizarre results. Referral arguments arise from a set of deductions from a body of knowledge and when this involves an inherent false premise results in an invalid argument. Referral arguments do not constitute an acceptable stand of proof.
In scientific methodology an "acceptable, statistical, certainty (high probability)" must exist as the accepted standard for a proffered fact to be accepted as evidence. However, in hermeneutics even a "probable" fact may be accepted in evidence by some because every possibility is relevant to the analysis. All relevant evidence necessary to give a complete picture remains imperative in either scientific or hermeneutic methods.
Thank you for visiting BIBARCH™