Given a piece of rock, any piece of rock, it is possible to determine something of its chemical or mineralogical composition, its structure and texture, and maybe even deduce something of its history. The procedures used in these studies will be more or less objective, and more or less precise (i.e. reproducible). They may involve non-destructive or destructive tests. The recorded data obtained will vary in quality, but the one certainty is that the rock itself will contain vastly more information than the scientist chose to extract from it. There is much more data that could be obtained from even the most intensively studied lunar samples.
Therefore, in one sense all geoscientific data can be said to be incomplete. The data should be recorded and stored in a database management system which allows for this incompleteness and also the imprecision of observational data. This is the subject of another paper in the conference.
There is a need for some way of estimating, expressing, and indicating the degree of subjectivity involved in interpreting data before they are recorded. For example, in assessing the crystallisation history of an igneous rock, the interpretation of the rock texture can be crucial, but this is a task which requires some skill and experience, and there can be more than one possible explanation for the same texture.
Is it possible to formalise a representation of 'degree of subjectivity' ? There is a whole range of data from the most objective (such as instrumental analyses) to the most subjective (such as description of rock texture). This is especially important as the same data type may often be measured in many different possible ways - for example 'colour' may be purely descriptive, or it may be supported by spectrometric determinations. Degree of uncertainty of a well-defined numeric measure can be expressed using well established statistical parameters. is there any similar treatment that can be applied to the degree of subjectivity ? Is it possible to quote some equivalent of 'standard error' for a colour description of 'dark-green' ?