Research meta-models
In the following sections, I outline some of the conceptual and methodological aspects (meta-models) that fuel and back-up my research. I hope these will some day make it into a proper publication, after some more thorough thinking.
This is a work in progress, in two senses:
- In one sense, it is not exhaustive and I will be adding content as it occurs to me.
- In the other sense, it is an ever-changing document, since I will frequently change my mind about any part(s) of content herein.
1 Research methods
As far as research methods are concerned, I am an enthusiastic advocate of Open Research, a term which encompasses “concepts of openness, transparency, rigor, reproducibility, replicability, and accumulation of knowledge” (Crüwell et al. 2019) put to the service of “making the content and process of producing evidence and claims transparent and accessible to others” (Munafò et al. 2017).1
I am also a supporter of the Slow Science movement, part of the broader “slow” movement (as in “slow-food”), which can be summarised with this quote from the manifesto: “We do need time to think. We do need time to digest. We do need time to misunderstand each other”.
Philosophy and science might seem two very unrelated fields, but in fact philosophy of science is an extremely important intersection of the two, with broad implications for research methods, from experiment planning, through statistical modelling, to inference. I’ll expand on this some time in the future, but for now I can strongly recommend the following books:
- Philosophy of science: Very short introduction (Okasha 2016).
- Understanding psychology as a science: An introduction to scientific and statistical inference (Dienes 2008).
- The Systems View of life: A unifying vision (Capra & Luisi 2018).
2 Phonetology
In a wish to transcend the phonetics/phonology divide (Sóskuthy 2013; Ohala 1990; Ohala 2005), I suggest the term phonetology (from a combination of phonetics and phonology) to refer to the study of what are traditionally investigated by phonology, phonetics and speech science.
The postulation of three levels of analysis (phonological, phonetic and motor-perceptual) is compatible with neuro-biological specialisations of the brain cortex (Basilakos et al. 2018; Fedorenko & Blank 2020; Ayyash et al. 2021; Fedorenko, Piantadosi & Gibson 2024; Fedorenko, Ivanova & Regev 2024; Regev et al. 2024). The tripartite nature of phonetological systems is reflected in the grammar architecture expounded in Bermúdez-Otero (2015) and Turk & Shattuck-Hufnagel (2020).
Although this newly-coined term contains the Greek root phon- ‘sound’, its meaning is by no means restricted to the realm of spoken languages, as its definition makes explicit. Sign languages, and other communicative systems yet to be “discovered”, are included. The term phonetological system can be used as an umbrella term that covers both visual and sound systems.
2.1 Phonetological systems
A phonetological system is comprised of a set of phonetological units and their relationships. Alas, I could not come up yet with a proper definition of “phonetological unit” that would subsume units of both visual and sound systems. Thoughts on this are much appreciated.
In my view of phonetology, a phonetological system has the following properties:
- It is a dynamic system (Thelen & Smith 2006; Vihman 2014).
- It is a physio-neuro-cognitive system. It has:
- A physical component (Ohala 1990; Ohala 2005).
- A neuro-cognitive component.
- It is rich in details (Johnson 1997; Pierrehumbert 2001; Bybee 2002).
- It is self-organising and emergent (Wedel 2007; Wedel 2011).
- It uses both stored exemplars and stored abstractions at multiple levels (Ambridge 2020).
- It is part of a complex socio-cultural system (de Boer 2015; Foulkes & Docherty 2006; Thompson, Kirby & Smith 2016).
I also consider these statements to be extendible more generally to Human Language as a holistic system.
3 Linguistic systems
In light of the brilliant ideas expounded in Cysouw & Good (2013) and van Rooy (2021), I believe linguistics would benefit from a formalised notion of linguistic system that would encompass any kind of system, from linguistic families down to idiolects, and any type of variation.
Cysouw & Good (2013) propose a meta-model based on linguistic documentation and description. From this model, I borrow two concepts:
Glossonym (p. 339)
A glossonym is a name (noun) used to refer to a linguistic system, whether a language, a lect, or a genealogically or areally related group of languages. A glossonym does not have a referent nor a significant, but rather it is just a signifier. For example, the glossonym Altaic, no actual reference is made to any linguistic entity, but only to the fact that this word has been used to refer to linguistic systems.
Doculect (p 342)
A doculect (i.e. documented lect) is a linguistic entity as it is documented/described in a specific resource (book, paper, booklet, website, etc.). This term is agnostic as to the type of linguistic entity (family, language, variety, etc.). A doculect is the pairing of a glossonym and a resource. An example: [Howard (1986); mawayana].
Cysouw & Good (2013: 347) also introduce the concept of “languoid” which refers to “any (possibly hierarchical) grouping of doculects, in principle ranging from a set of idiolects to a high-level language family”.
I propose instead the term glossolect, which—unlike languoid—does not necessarily entail a strictly-nested hierarchical structure. This difference accommodates those cases where linguistic grouping is not based on clear-cut splits, but rather based on a network of shared features (cases which are possibly more frequent), in which a doculect can in fact belong to different intersecting glossolects.
A glossolect is thus a set of doculects and can refer to any of a macro-family, a family, a dialect continuum, a linkage, a language, a variety, an idiolect, etc.
In terms of types of linguistic variation, I borrow from classical sociolinguistics (references TBA):
- Diachronic: variation in time.
- Diatopic: variation in geographical space (less frequently used is diachoric for more restricted spatial variation).
- Diastratic: variation based on social class and groups.
- Diaphasic: variation based on linguistic and social context (includes registers and slangs).
- Diamesic: variation based on the means of communication (e.g., visual/oral vs written, web vs print, etc.).
In sum:
- A glossolect is a set of doculects.
- A doculect is a pairing of a glossonym and a resource.
- A glossonym is a name used to refer to a linguistic entity.
- A resource is any object (publication, media, etc.) containing linguistic materials.
- Glossolects can be situated within a specific diachronic, diatopic, diastratic, diaphasic, diamesic context.
4 Linguistics
The following sections provide a thought-out description of my ontological and epistemological beliefs in relation to “linguistics” and its object(s) of study. The reader should interpret all of the statements made here as “my personal opinion”. In some cases opinion refers to my expert opinion, in others to my intuitions (for the use of intuition in science, see the work by physicist Fritjof Capra). With time, I will add more references where/if relevant.
4.1 Ontological
- Human Language is the human ability to communicate through a language, and a language is a specific communicative system used by a specific group of humans.
- linguisticality is the prerequisite for Human Language (Haspelmath 2020).
- If one defines Human Language as the human ability to communicate through a language, then saying that only humans have Human Language is a tautology.
- All living beings are capable of communicating (even plants and fungi, Mancuso & Viola 2015).
4.2 Epistemological
- Linguistics is usually defined as a “science that studies Human Language”, however there is no agreement as to what science actually is.
- Science is not the only way to obtain knowledge. There are many epistemological practices or “ways of knowing”. Like all ways of knowing, science is human activity.
- I prefer to use the term research instead of science for all purposes.
- Research has both objective and subjective components.
- Statistics and statistical inference have both objective and subjective components (Gelman & Hennig 2017).
- There is no agreement as to what language, Human Language, linguisticality actually are.
- There is no agreement as to whether language, Human Language, linguisticality are the same thing or if they in principle refer to different entities.
- There is no agreement as to which components of the human ability to communicate through a language are innate and which are learned.
- There is no agreement as to what innate actually means.
- There is no agreement as to whether the human ability to communicate through a language is entirely domain-specific or entirely domain-general, or if it is a bit of both.
- If it is a bit of both, there is no agreement as to which components are domain-specific and which are domain-general.
- There is no agreement as to what a language is. In fact, in English language is used at times to mean either the human ability/capacity to communicate with a language or to mean a specific language (sometimes written Language and language respectively). See Section 3.
References
Footnotes
Open Research is a more inclusive term than the more common Open Science term.↩︎