We agree with Frost that the variety of orthographies in the world's languages complicates the task of “cracking the orthographic code.” Frost suggests that orthographic processing must therefore differ between orthographies. We suggest that the same basic orthographic processes are applied to all languages. Where languages differ is in what the reader must do with the results of orthographic processing.

Reading speed matters in most real-world contexts, and it is a robust and easy aspect of reading to measure. Theories of reading should account for speed.

We qualify Frost's proposals regarding letter-position coding in visual word recognition and the universal model of reading. First, we show that perceptual uncertainty regarding letter position is not tied to European languages – instead it is a general property of the cognitive system. Second, we argue that a universal model of reading should incorporate a developmental view of the reading process.

Languages may get the writing system they deserve or merely a writing system they can live with – adaption without optimization. A universal theory of reading reflects the general dependence of writing on language and the adaptations required by the demands of specific languages and their written forms. The theory also can be informed by research that addresses a specific language and orthography, gaining universality through demonstrating adaptations to language and writing input.

We agree with many of the principles proposed by Frost but highlight crucial caveats and report research findings that challenge several assertions made in the target article. We discuss the roles that visual processing, development, and bilingualism play in visual word recognition and reading. These are overlooked in all current models, but are fundamental to any universal model of reading.

The search for a universal theory of reading is misguided. Instead, theories should articulate general principles of neural computation that interact with language-specific learning environments to explain the full diversity of observed reading-related phenomena across the world's languages.

We applaud Ram Frost for highlighting the need for multicultural perspectives while developing universal models of visual word recognition. We second Frost's proposal that factors like lexical morphology should be incorporated besides purely orthographic features in modeling word recognition. In support, we provide fresh evidence from Hindi (written in Devanagari), an example of hitherto under-represented alphasyllabic orthographies, in which flexible encoding of akṣara (character) position is constrained by the morphological structure of words.

One key insight of Frost's target article is that morphology has priority over phonology in writing and in cognitive processing. I argue that this insight raises challenges for theories that put phonology at the heart of the reading process. Instead, it highlights the potential importance of a morphemically based visual pathway to meaning in this process.

A spelling model which has evolved in the parallel universe of spelling research resonates with Frost's reading model. Like reading, spelling cannot be based solely on phonology or orthography, but should accommodate all linguistic facets. The cognitive domain of spelling does not take place at the level of single grapheme or phoneme or syllable, but rather, at the lexical level.

Frost's critique reveals the limitations of the reverse-engineering approach to cognitive modeling – the style of psychological explanation in which a stipulated internal organization (in the form of a computational mechanism) explains a relatively narrow set of phenomena. An alternative is to view organization as both the explanation for some phenomena and a phenomenon to be explained. This move poses new and interesting theoretical challenges for theories of word reading.

Languages and writing systems result from satisfying multiple constraints related to learning, comprehension, production, and their biological bases. Orthographies are not optimal because these constraints often conflict, with further deviations due to accidents of history and geography. Things tend to even out because writing systems and the languages they represent exhibit systematic trade-offs between orthographic depth and morphological complexity.

I argue that the study of variability rather than invariance should head the reading research agenda, and that strong claims of orthographic “optimality” are unwarranted. I also expand briefly on Frost's assertion that an efficient orthography must represent sound and meaning, by considering writing systems as dual-purpose devices that must provide decipherability for novice readers and automatizability for the expert.

Letter-position tolerance varies across languages. This observation suggests that the neural code for letter strings may also be subtly different. Although language-specific models remain useful, we should endeavor to develop a universal model of reading acquisition which incorporates crucial neurobiological constraints. Such a model, through a progressive internalization of phonological and lexical regularities, could perhaps converge onto the language-specific properties outlined by Frost.

Interest in orthographic processing reflects an expansion, not constriction, of the scope of research in visual word recognition (VWR). Transposition effects are merely one aspect of investigations into orthographic encoding, while open bigrams can accommodate differences across languages. The target article's inaccurate characterization of the study of orthographic processing is not conducive to the advancement of VWR research.

I have argued that orthographic processing cannot be understood and modeled without considering the manner in which orthographic structure represents phonological, semantic, and morphological information in a given writing system. A reading theory, therefore, must be a theory of the interaction of the reader with his/her linguistic environment. This outlines a novel approach to studying and modeling visual word recognition, an approach that focuses on the common cognitive principles involved in processing printed words across different writing systems. These claims were challenged by several commentaries that contested the merits of my general theoretical agenda, the relevance of the evolution of writing systems, and the plausibility of finding commonalities in reading across orthographies. Other commentaries extended the scope of the debate by bringing into the discussion additional perspectives. My response addresses all these issues. By considering the constraints of neurobiology on modeling reading, developmental data, and a large scope of cross-linguistic evidence, I argue that front-end implementations of orthographic processing that do not stem from a comprehensive theory of the complex information conveyed by writing systems do not present a viable approach for understanding reading. The common principles by which writing systems have evolved to represent orthographic, phonological, and semantic information in a language reveal the critical distributional characteristics of orthographic structure that govern reading behavior. Models of reading should thus be learning models, primarily constrained by cross-linguistic developmental evidence that describes how the statistical properties of writing systems shape the characteristics of orthographic processing. When this approach is adopted, a universal model of reading is possible.

The science of genetics is undergoing a paradigm shift. Recent discoveries, including the activity of retrotransposons, the extent of copy number variations, somatic and chromosomal mosaicism, and the nature of the epigenome as a regulator of DNA expressivity, are challenging a series of dogmas concerning the nature of the genome and the relationship between genotype and phenotype. According to three widely held dogmas, DNA is the unchanging template of heredity, is identical in all the cells and tissues of the body, and is the sole agent of inheritance. Rather than being an unchanging template, DNA appears subject to a good deal of environmentally induced change. Instead of identical DNA in all the cells of the body, somatic mosaicism appears to be the normal human condition. And DNA can no longer be considered the sole agent of inheritance. We now know that the epigenome, which regulates gene expressivity, can be inherited via the germline. These developments are particularly significant for behavior genetics for at least three reasons: First, epigenetic regulation, DNA variability, and somatic mosaicism appear to be particularly prevalent in the human brain and probably are involved in much of human behavior; second, they have important implications for the validity of heritability and gene association studies, the methodologies that largely define the discipline of behavior genetics; and third, they appear to play a critical role in development during the perinatal period and, in particular, in enabling phenotypic plasticity in offspring. I examine one of the central claims to emerge from the use of heritability studies in the behavioral sciences, the principle of minimal shared maternal effects, in light of the growing awareness that the maternal perinatal environment is a critical venue for the exercise of adaptive phenotypic plasticity. This consideration has important implications for both developmental and evolutionary biology.

A number of recent developments in our understanding of the biology of heritability question commonly held views on the immutability of genetic factors. These have numerous potential implications for improving understanding and practice in pre- and postconceptional care and for infant and child mental health, and they carry a cautionary message against overgeneralization.

Several new molecular findings and concepts furnish evidence in support of gene–environment interdependence, challenging some of the current tenets and basic statistics of behavioral genetics. I, however, argue that (1) some of the expectations evoked by “neogenomics” are contradicted by findings; and (2) while epigenetic and gene expression effects are complex, they can to some extent be incorporated into “classical” behavioral genetics modeling.

Charney argues that the presence of inherited epigenetic effects makes twin, family, and adoption studies obsolete. This argument relies on both a faulty characterization of these studies and indirect comparisons of DNA and “neogenetic” factors. I argue that twin and family studies will in fact serve a necessary and vital role in the study of epigenetic and neogenetic processes.

Charney describes several mechanisms that will bias estimates of heritability in unpredictable directions. In addition, the mechanisms described by Charney explain the puzzling fact that research in human-behavior genetics routinely reports higher heritabilities than animal studies do. However, I argue that the concept of heritability has no real place in human research anyway.