Theory in plant science

Authors


Conversations with experimentalists about theory are often very short, with no real engagement about what is quite central to the scientific method. An objective approach for detecting the presence of theory in published papers can be achieved by searching academic databases. The term ‘theory’ was used to search all publications in New Phytologist for 2009 and 2010, using Thomson ISI Web of Knowledge, SciVerse Scopus, Google Scholar and the Wiley Online Library. The search engines produced markedly different quantitative information; ISI detected 12 papers, SciVerse Scopus 102, Google Scholar 193 and Wiley Online Library 362. The citations were then normalized and allocated to the New Phytologist sections: Physiology and Development, Environment, Interaction and Evolution (Fig. 1).

Figure 1.

Normalized citation frequency of ‘theory’ in New Phytologist between 2009 and 2010, using four search engines and allocated to the New Phytologist sections: Physiology and Development (Phys/Dev), Environment, Interaction and Evolution.

In spite of widely differing total citations from the different search engines, the overall conclusion is that theory figures strongly in papers in the Evolution and Environment sections of New Phytologist and less so in Physiology and Development and in Interaction.

The development of plant science is based on observations, the development of theories to explain these observations and the testing of these theories. ISI will count 664 papers between 2009 and 2010 in the calculation of the 2011 Impact Factor. Does this mean that only 2% of papers, according to ISI, or a maximum of 55%, according to Wiley, report on science that has been pursued in the accepted approach? This seems highly unlikely and is clearly a measure of the wide differences in search routines of the search engines. In fact science develops by two different approaches (Fig. 2).

Figure 2.

Development of science by induction and deduction.

Deductive scientific development in plant science can be illustrated by the theory of evolution by natural selection, derived by Darwin. This is an all encompassing theory and needs to be reduced in scope to hypotheses but enhanced in detail to be operational. A sub-component of evolution by natural selection is hybrid zone theory, or in terms of Fig. 2 the hybrid zone hypothesis (Brennan et al., 2009). This addresses the mechanisms that maintain the zone. Observations in these zones can be used to quantify how they are maintained in a situation of selection and gene flow. New observations may lead to changes in the hypothesis but are unlikely to change the overall theory of natural selection.

Darwin developed his theory of evolution inductively, based on large collections of observations, moving through smaller scale hypotheses to the overall large picture theory. This theory has since been a continual source of hypotheses for testing the validity and consequences of the theory.

The breadth of papers published in New Phytologist provides a source of information to determine whether deductive or inductive approaches dominate overall and in the different sections of the journal. Overall, papers can be classified in both approaches and in all sections. Papers classified as deductive are those that have an overall banner of theory that drives, in general, the development of more specific hypotheses to be investigated. Papers classified as inductive are those that collect and use data sets to derive hypotheses about how a process or mechanism operates. The deductive approach dominates in the Evolution section. Although these papers can all consider natural selection as the overall theory, particular hypotheses need to be developed for testing and are quite distinct, from the importance of pollen grain selection (Hasegawa et al., 2009) and floral polymorphism (Dormont et al., 2009) in seed production, through quantification of crop transgene persistence (Snow et al., 2010), to identifying major problems in estimating the dating of polyploid events (Doyle & Egan, 2010).

Papers in the Environment section are predominantly inductive, although Price et al. (2010) provide an unusual example of trying to establish a new theory. More typical are approaches to accumulate data sets of observations, to provide hypotheses of environmental controls, such as the impacts of precipitation events on desert productivity (Robertson et al., 2009) and comparing differences in species characteristics, such as recovery from stem damage in tree species (Romero et al., 2009) and variations in root traits (Comas & Eissenstat, 2009). The application of molecular approaches in ecology is increasing and provides a powerful approach to explain the current distribution of vegetation types in terms of past geological events (Qiu et al., 2009). Molecular approaches are central to many papers in the Physiology and Development section and Rowan et al. (2009) use the approach inductively in exploring the environmental regulation of anthocyanin biosynthesis. Deductive approaches investigate proposed mechanisms, such as the role of NAD metabolism in seed dormancy (Hunt & Gray, 2009) and the role of phytochrome in a non-photosynthetic plant parasite (Takagi et al., 2009). One of the purest examples of rejecting a theory is by Duckett et al. (2009), who clearly demonstrate that capsule dehiscence in Sphagnum is not due to the build up of air pressure as the capsule dries out, a theory of significant longevity, but is in fact due to differential shrinkage of the capsule walls.

The Interaction section has papers with both deductive and inductive approaches to research. The interactions between fungal partners and their host plants are typical avenues of deductive enquiry, such as the mycorrhizal uptake of nitrogen from organic material (Leigh et al., 2009), the impact of plant defence compounds on non-mycorrhizal fungal endophytes (Saunders & Kohn, 2009) and the cryptic nature of the orchid mutualism (Cameron et al., 2009). However, the episodic nature of larch bud moth outbreaks (Büntgen et al., 2009) requires long-term data sets and an inductive approach to investigate the nature of this pseudo-cyclic interaction.

New Phytologist author guidelines state clearly that submissions should be original research that addresses clear hypotheses and offers new insights. In Fig. 2 hypotheses are central to both deductive and inductive approaches to science. Neither approach has an ascendancy over the other in plant science – but perhaps après moi le déluge !