John Bateman On What Is And Isn't An Instantiation Of Theory

John Bateman replied to David Rose on sysfling list at 13 May 2018 at 21:06:

Now, observations are not instantiations of theory; they cannot be, otherwise we could only observe what our theory says. And thus we could not observe anything which did not conform to theory: this may be the case with certain kinds of discourses that call themselves theory, but I have problems with those. [another topic!] 

I would suggest (discussion sought) that instantiations of theory are *hypotheses*, that is, they are descriptions of how the world would be given the theory. 

Observations are then distinct from that, since they aim to listen to how the world actually is rather than how the theory says it is. 

Observations are, however, never divorced from theory, since theory gives us the basis for making observations: the relation just cannot be (simply) one of instantiation. 

So, observations must be partial instantiations of theory in the sense that they are what you get when you take some phenomena and construe those phenomena drawing on some partial instantiation of theory. The result is a description that relies on the theory without being a full instantiation (like a partially frozen idiom). Only then is it possible to find places where there is a lack of fit between what the theory predicts (its instantiations) and what was observed (construals/descriptions of data where the descriptions are partial instantiations of theory). One might be willing to drop certain parts of the theory more readily than others. This is then the practice of exploring and developing theory. …

This also provides further methodological and theoretical support for pursuing triangulation: since one can use descriptions of phenomena partially instantiating different theories to play off against one another. Single theories might not have sufficient resources for doing that. 

And back to David's:

His ‘weather and climate’ analogy for instantiation is an apt metaphor for the observation/theory relation in SF research.

I'd be hesitant about that, exactly because the relationship between observation and theory is not instantiation. So you can't get quite the same time-depth phenomena... 

Blogger Comments:

[1] On the question as to whether the relation between observation and theory is one of instantiation, Bateman writes:

• observations are not instantiations of theory
• the relation just cannot be (simply) one of instantiation
• observations must be partial instantiations of theory
• the relationship between observation and theory is not instantiation.

To be clear, a "partial instantiation" is instantiation, and so this is self-contradiction.  On the SFL model, an observation is an instance of observation potential, and an instance of theory potential is a theory.

[2] On the question as to what constitutes an instance of theory, Bateman writes:

• instantiations of theory are *hypotheses* 
• descriptions of how the world would be given the theory 
• observations must be partial instantiations of theory 
• what you get when you take some phenomena and construe those phenomena drawing on some partial instantiation of theory 
• a description that relies on the theory without being a full instantiation 
• what the theory predicts 
• what was observed (construals/descriptions of data where the descriptions are partial instantiations of theory) 
• descriptions of phenomena partially instantiating different theories.

To be clear, Bateman identifies instances of theory as theoretical predictions and descriptions of phenomena/data, the latter equated with both hypotheses and observations.  Even ignoring the differences between predictions, descriptions, hypotheses and observations, this contradicts his own claims that

• observations are not instantiations of theory, and
• observations are distinct from hypotheses/descriptions.

[3] To be clear, on the SFL model, "how the world actually is" is itself a construal of experience as meaning, either through language (first-order meaning), or through the re-construals of linguistic meaning as theory (second-order meaning).