(ii.2) Creating affect can be used for narrative interaction in the classroom or with the public—making teachers good storytellers (Kubli, 2001, 2005).

(ii.3) Last but not least, stories are a perfect environment for creating historical and social context for science (Klassen, 2006; Metz et al., 2007; Levinson, 2006).

Narrative for science

Different forms of narratives have been used for supporting science content, learning of science and science method, and formulation of beliefs about science.

With this we come closer to an intrinsic relationship between science content and narrative. Still, there remains a considerable gap between the use of narrative for supporting science and the full-fledged claim that science can be intrinsically narrative (see the next two subsections for a discussion of Subcategory iv). The distinction between narratives being either intrinsic or extrinsic to science, has been made by Norris et al. (2005).

(iii.1) A first application in Subcategory iii consists in blending expository (descriptive) text with narrative elements that are either extrinsic (as in ii or iii) or intrinsic (iv). It creates a narrative setting (ex-tended oral or written accounts) for conveying accepted scientific terms and concepts (see Avraamidou and Osborn, 2009). This leads to results similar to iii.2.

(iii.2) Secondly, narrative for supporting science learning has been explored for environments that try to be conducive to conceptual change. Klassen (2010), in particular, argues that the progression of learning resembles a story—making narrative environments the right setting for productive learning, including explicit conceptual change. This is achieved by relating models narratively to accepted theory.

(iii.3) The third application in this category is very interesting. Ochs et al. (1992) have found by studying verbal exchanges between adults and children at the dinner table that critical thinking and scientific perspectives are formed in narratively shaped discourse. This makes it possible to use narrative to expose learners to argument/dispute as a preparation for nature of science issues; it provides training for forms of reasoning and learning to understand perspective. Finally, Bruner’s observation (1996, p. 126) that the process of science (hypothesis formation, testing, correcting, etc.) is narrative fits neatly in this category: storytelling supports scientific thinking in general.

(iii.4) Finally, there is the counterpart to science as narrative (i.1: science is a narrative about who we are). Stories are a perfect environment for a narrative creation of a web of scientific beliefs. This is the same as taking a particular philosophic stance—deciding how we can or want to understand the world from a scientific perspective. Formal science cannot do this, that is not its job; the kind of thinking needed to form scientific beliefs is by nature narrative (Herman, 2009, Chapter 5).

Narrative as science: Narrative intrinsic to science

Let me now discuss what may rightly be called narratives intrinsic to science (Norris et al., 2005). There is research that has made important contributions to narrative explanation in science and narrative use of models which I would like to discuss here. The third form of narrative in this category, narrative framing (suggesting of concepts and models) will be described in the following subsection.

(iv.1) Norris et al. (2005) have written an extensive survey of forms of narrative in science, both extrinsic and intrinsic. This makes it possible for me to be brief—in fact, their research is a valuable source of background information for what I have only touched upon above; they recount in some detail how we might want to define narrative. With regard to the phenomenon of explanation, the authors argue against a narrow reading of explanation only in terms of the deductive-nomological model (see, for example, Hempel and Oppenheim, 1984). Norris et al. (2005) show how the deductive reading of explanation cannot do justice to a vast range of science. Where science treats either singular events (a meteorite striking the Earth 65 million years ago) or historical events in nature (evolution, the development of a particular ecosystem), strict deduction fails. In the end, only narratives of the (special, singular, historical) events can be produced. Such stories properly and sufficiently explain what we want to know.

The application of explanatory narrative in historical natural science—when nature has a history—has since been widely accepted. Philosophers of earth science are among those who explain to us how their field can be scientific and narrative at the same time (Kleinhans et al., 2010). It is no surprise that the issue of narrative explanation is a hot (and hotly contested) subject in the study of history (Carr, 2008; Velleman, 2003), and researchers in the historical natural sciences profit from this debate.

A particularly interesting case of science as historicizing narrative has been described by Wise (2011). Wise contrasts the traditional mode of explanation as deduction from differential equations with explanation through simulation. He argues that narratives that accompany simulations are natural historical in kind: they (the narratives) explain phenomena by growing (developing) them rather than by referring them to general laws. This is indeed a phenomenon well known to those working in computational fields of science: explanations (of the behavior of systems) grow from many simulations; a picture emerges from a vast number of trajectories followed rather than from a single analytic solution of a set of equations.

(iv.2) The following application of narrative in (social) science has been described in a noteworthy publication by Mary Morgan (2001). I consider this paper on economic modeling one of the most important (not the least for the natural sciences) for showing how far the theory of narrative in science has come. Morgan demonstrates that using models and relating them to the world is a narrative activity. How to make use of a model or what to look for in a simulation are aspects that are not covered by the model or its underlying theory. Posing questions that lead to the definition of parameters and initial values for subsequent simulation is a narrative act.

In summary, simulating a model is telling a story. In the case of dynamical models, telling these stories means creating a history of and for the system being modeled; Morgan’s example of the application of narrative in science is definitely related to iv.1, but her emphasis of how to embed models in the world is worth discussing separately. (See also McCloskey, 1990.)

Morgan’s (2001) discussion is important in another respect as well. It shows where researchers seem to draw a line when it comes to using narrative as science. Morgan makes it clear that storytelling means simulating (existing) models that were derived from theory in classical manner. Theory is given—it exists prior to modeling and the subsequent narrative act of embedding models in the world. Storytelling does not seem to lead to theory. [This is similar to another important contribution to our theme of explanation: Ohlsson (1992) speaks of theory articulation, i.e., application of theory that must take narrative forms since the procedures to apply a theory are not part of the theory itself. Again, Ohlsson does not suggest the reverse, that using narrative would lead to theory. Closer reading, however, shows that Morgan speaks of using stories for suggesting possible extensions and changes to existing models. This is a step in the direction of framing and conceptualization.]