The world’s leading publication for data science, AI, and ML professionals.

Between Language and Furniture

Do you think there are any commonalities?

Image by Author
Image by Author

Couple of years ago I was reading about those two seemingly different topics, and I was surprised by the connections I made, or maybe I was too bored back then and had nothing else to do, so I hallucinated some connections!

The following threads of thought were in a sense the precursor for my Master of Science thesis in Architectural Computation at The Bartlett – University College London in 2017. The thesis explored the viability of using Machine Learning and Open Big Data to automate design tasks.

Following, I highlight the importance of furniture, then draw similarities between furniture layout schemes and natural language systems. Claiming that these similarities between the two can aid in tackling the furniture layout planning (FLP) problem from a different perspective.

Furniture Layout Planning or FLP is a problem that is concerned with finding optimal organisations for furniture within a given space.

The post then concludes with a discussion, proposing how to expand on the presented ideas, and further development possibilities by applying solutions used within the natural language processing domain.


Introduction

One of the tasks that consumes a lot of time during the design development phase, of any architectural project, is generating valid furniture layout schemes. A lot of studies explored different techniques for generating spatial configurations, this is done under the topic of space layout planning. Few studies focus on the generation of furniture schemes. Generating such schemes is an essential step in the design process of any functional space. Most of the current FLP research utilises exhaustive search algorithms which produces mass number of schemes, most of which are illogical, as concluded by Abdelmohsen et al (2017).

Similarities will be drawn between natural language systems and the FLP problem, trying to investigate the existence of underlying structures and rules governing the distribution of furniture elements within a space. The post is intended to open a wider framework of study that tackles the FLP problem from a different perspective.

I claim that, studying the rules regulating the configuration of furniture elements inside a space, would act as the basis for further research. Research that should solely focus on testing the generative nature of such rules. Finding those similarities between natural languages and FLP, would allow the realisation of a more informed generative schemes, that does not solely depend on exhaustive search approaches. It is essential to test the applicability of using such rules, in the development of tools capable of solving the FLP problem for any given space. This would result in circumventing the amount of time lost on the FLP task in a professional context.

Furniture as a Language

The definition of language that I am interested in is language as a "system of distinct signs corresponding to distinct ideas" (Saussure, 1959).

Saussure’s definition, does not limit language to that of the spoken, but rather allows the term to encompass gestural methods of communication. It even opens the possibility for discussing FLP as a language. The signs in the case of FLP is considered the distinct furniture elements.

The idea of dealing with furniture items as signifiers, could be made more clear if we look at a Charles Eames’ Lounge chair and a casual bar stool, both elements acts as signifiers for different meanings, each signifies different values of comfort, luxury, expected time of use or even modes of use.

Eames Lounge Chair and Ottoman © HermanMiller.com All Rights Reserved
Eames Lounge Chair and Ottoman © HermanMiller.com All Rights Reserved

Not only are the similarities to language systems evident when looking at furniture pieces as standalone elements, but also when taking into consideration the relationship between multiple elements and their respective relationship to the surrounding space. Those two relationships, element to element and element to space, brings up another analogy, which is language as syntax and semantics (Chomsky, 2002).

The syntactic and semantic aspect of furniture elements within the space are explained briefly in (Figure 1).

Figure 1. a) Shows a grammatically incorrect position for the sofa within the space, while b) shows a grammatically correct positioning of elements but not necessarily semantically valid.
Figure 1. a) Shows a grammatically incorrect position for the sofa within the space, while b) shows a grammatically correct positioning of elements but not necessarily semantically valid.

Following this analogy, we can look at furniture pieces within a space as a collection of distinct entities, each could be considered, like words, a signifier for a specific meaning. This collection of entities, like words in any language, can be organised into infinitely different schemes within a space. This "corpus" of infinite organisations, like language, can be categorised into grammatically correct and semantically sensible schemes, grammatically correct but semantically nonsensical schemes and lastly not grammatical nor semantically sensible schemes.

Furniture and Dwellings

In Building Dwelling Thinking, Heidegger (2012) speaks of dwellings, he explains it as the place where one feels at home, and that does not necessarily mean one’s house, as he gives an example "The truck driver is at home on the highway".

I can see this notion of dwelling presented on different scales, starting from one’s house, to one’s neighbourhood, to the city and so on. Another interesting notion in this context, is appropriation and how being is linked to the capacity to appropriate (Malpas, 2006). I claim that, one of the methods by which people appropriate their surrounding space, and by which space presents affordance for appropriation, is furniture.

We interact daily with furniture in our houses, in public transport, at work or college and during leisure time at a restaurant or a bar. Furniture is an integral part in our daily routines and it influences how we interact with the space around us and with one another.

Take for example two similar studies, one recorded what people do during their trip on the underground tube in London and another recorded what people do during their rail trips in London as well.

The first found that the most utilised activity during underground trips is technology usage (Gamberini et al., 2013). While the other found that the favourite activity during rail trips is window gazing/people watching (Lyons, Jain and Holley, 2007).

If we disregard the difference between both,in respect to having a scenery to gaze at or not, one thing both studies did not take into consideration is the seating layout, Figure 2.

Figure 2. LEFT. Shows seating layout for London's underground tube (PA, 2014), while RIGHT. shows seating layout for London's rail service (Chris Radburn/PA Wire, 2014) © All Rights Reserved to Respective Authors
Figure 2. LEFT. Shows seating layout for London’s underground tube (PA, 2014), while RIGHT. shows seating layout for London’s rail service (Chris Radburn/PA Wire, 2014) © All Rights Reserved to Respective Authors

It could be noticed that in the underground, there exists "conflicting feelings of togetherness and distance which passengers experience, the activity of ‘waiting’ and exposure to ‘other-ness’" (Bassoli and Martin, 2006).

One of the ways afforded by the seating layout in the underground is resorting to the "absent presence", where "one is physically present but is absorbed by a technologically mediated world of elsewhere" (Gergen, 2002), watching people or just gazing might be unsettling for others.

While the layout on the national rail, affords people to watch others and not being spotted, or at least thinking they are not, in the comfort of their own field of views, which only includes the back of people heads.

This begs further research, but at least it simply shows how furniture layout might direct people into certain modes of interaction.

Seven Functions of Furniture

In his book ‘Famous Last Words’ Daniels (1983) builds on Halliday’s (1977) discovery of the seven main functions of language, and the similarities are glaringly obvious when applied to furniture as well.

He says language is "instrumental, regulatory, interactional, personal, heuristic, imaginative and representational".

Instrumental, to get things done, in the furniture domain this is evident in studies that show how the ergonomics of office design increases the efficiency of employees (Hammed and Amjad, 2009).

Regulatory, trying to control how other people act, in furniture domain this is evident in the layout used for a lecture for example, putting the focus of the listeners on the person standing in front of them.

Interactional, for the definition of groups and relationships, this could be seen in informal furniture layouts utilised in co-working spaces and startups to enhance communication, extraversion, and collaboration. To tie back briefly to the syntax and semantics part, if interaction is the main goal, furnishing a collaboration space like a lecture room space would be considered grammatically correct yet semantically not sensible.

Personal, expressing oneself and conveying certain feelings, believes or statements, which is similar to how people consciously or unconsciously choose the furniture pieces for their own home.

Heuristic, to learn or test a hypothesis, similarly to the example given in the regulatory aspect, such a study to be performed entailed testing out different furniture pieces with different organisations to see how people would start acting in a certain space.

Imaginative, how language is used to create fiction, imagine how a child would orient his chair or elevate it on a table in front a group of his friends to play a king’s character.

Representational, where language is used to provide information or communicate a specific matter, imagine you are going into your classroom and you find the chairs laid out consecutively facing the projector, and imagine on the next day you found them distributed into clusters, this conveys information about how your day will go, listening to a lecture or doing group work/exercise.

The examples for each one of these points vary, but in a general sense this sheds a light on the similarities between language and furniture.

Innateness

In ‘Syntactic Structures’ Chomsky (2002) discusses the idea of how the human mind is capable of learning new languages, but not by going through all the utterance of said language, which is infinite, but rather through an innate faculty that allows us to discern languages at young age and stat speaking it, he uses this to point out the underlying generative nature of language.

This innateness could be explained better by how Hillier (1998) draws similarities between our mind’s ability to unconsciously handle configurations and the way we manage to do the same for syntactic and semantic structures of languages.

It could be said then that this innate capability is not only for languages but rather for complex relational schemes in general. And like language, furnishing a space is something many people could do, without going through a methodological educational process to gain it as a skill. People manage to draw relationships between, needs, traditions, believes and furniture layout inside their households.

For example, in most middle eastern households, women do not usually resort to the use of a professional interior designer or an architect to help them furnish their homes. Yet there are certain traits that exists in most households without being explicitly taught. For example, due to religious and traditional believes, a visitor shouldn’t see the women in the house if they are not veiled. Not only does this criterion guide the furnishing of the public zone of the house, but even if it is not accomplished in a good manner, it would dictate how the visitors would choose their seating spot, Figure 3.

This shows how a higher level of design intent adds a semantic aspect to the furniture scheme.

Figure 3. a) Would be a semantically incorrect option, while both b) and c) would be considered semantically sensible, all though are considered grammatically correct.
Figure 3. a) Would be a semantically incorrect option, while both b) and c) would be considered semantically sensible, all though are considered grammatically correct.

Furniture as a dialect

One of the interesting similarities between language and furniture is the notion of a dialect.

Dialect is the variation that happens to a language and affects its vocabulary, pronunciation, and Grammar while most time conserving the semantics. These variations can occur due to "regional, social or stylistic changes" (Rickford, 2002).

Take for example, the difference between a modern living room and a middle-eastern majlis, Figure 4. Semantically both have the same meaning, a welcoming space for guests, but the grammar changes due to different traditions, for example the layout for the majlis must always have a focal point, which is where the owner of the house is seated, also different circulation constrains governs the element to element and element to space relationships, for example more circulation space is required in front of the furniture for workers to serve the traditional drinks to the guests as well as to give space for people to get up or sit on the floor easily.

Figure 4. a) Shows a middle-eastern majlis (ouweland-akhmad, 2017), b) Shows a modern living room (GM Developments - Michael Maynard, 2017)
Figure 4. a) Shows a middle-eastern majlis (ouweland-akhmad, 2017), b) Shows a modern living room (GM Developments – Michael Maynard, 2017)

Furniture Semantics

The notion of semantics in furniture layout schemes has been touched upon during some of the previous examples. But to make it more clear, when we talk about semantics in that context, we are referring to a higher level of design intent. Not just simple considerations like the furniture shouldn’t block door access, or the furniture shouldn’t be placed facing the wall and not providing access to users. When we speak about higher level design intent, we mean things like how the layout should have visual balance within the space, or how it enhances the interaction or conversations between people, how the layout can emphasise certain aspects of the space (Merrell et al., 2011) or how, as mentioned previously, can it have religious, regional, class or even artistic connotations.

Discussion

I have presented various ways by which one can see the similarities between the structure of natural languages and the use of furniture in a space. These similarities were drawn to pave the road for figuring out if techniques used in natural Language processing could be utilised in the FLP domain.

Could NLP applications like spelling and grammar checking be used for analysis of furniture schemes, optical character recognition and information extraction for extracting useful insight from mass number of plans available online, or using machine learning, can we use question and intent analysis to start proposing schemes.

Leveraging such tools in this domain would be helpful in the context of the design process, and also for facility layout and space planning and management, it might be even useful for the general public.

With the rise of autonomous and robotically controlled systems, one might imagine that such an understanding of the underlying rules of furniture planning coupled with projections of how people act in the space and an awareness of the current usage intent, along with the availability of adaptive kinetic furniture systems might not be an over reach.


Related Articles