You are on page 1of 27

JOURNAL

OF

ANTHROPOLOGICAL

ARCHAEOLOGY

2, 277-303

(1983)

A Theory of Architectural

Design

 

RANDALL

H. MCGUIRE

Department

of Anthropology,

Binghamton,

State University

of New

York,

New York 13901

AND

MICHAEL

B.

SCHIFFER

Department

of Anthropology,

University

of Arizona,

Tucson, Arizona

85721

 

Received January

5, 1983

A rudimentary

sented. Conceiving

theory

to explain

of architectural

the design of vernacular

structures

is pre-

design as a social process,

the theory

focuses

on the influence

of utilitarian

and symbolic

functions

as well

as on the trade-offs

between production and

maintenance

costs. A particular

design is viewed

as the

outcome

of a process

of

compromise

among conflicting

goals, influenced

by

fac-

tors

of adaptation

and social

organization.

The theory

is used

to generate

an

explanatory

sketch for

why the prehistoric

from being pithouse

Anasazi of the American Southwest

went

to pueblo dwellers.

INTRODUCTION

The purpose

of this paper

is to advance

a preliminary

but general theory

to

explain

the design

of vernacular

architecture.

This effort

is intended

to contribute

to

the larger

research

emphasis

developing

in archaeology

that

is concerned

with

explaining,

 

in behavioral

terms,

variability

and

change

in material

culture

(e.g., Braun

1983; Schiffer

1979; Hayden

1977a,

1977b;

Goodyear

1979).

Although

the

determinants

of

specific

artifact

morphologies

have

always

been

of

interest

in archaeology,

theoretical

treatment

of

the design

process

is usually

subsumed

by discussions

of

“style”

and “function.”

These

discussions

have resulted

in tangible

progress

(e.g.,

Wobst

1977; Plog

1980; Dunnell

1978; McGuire

1981;

Sackett

1982; Rathje

and Schiffer

1982; Jelinek

1976), but

a fully

general

theory,

applicable

to architecture

as well

as to chipped-stone

tools,

re-

mains elusive.

The present

paper

may contribute

to the construction

of

a high-level

theory

of artifact

design, but our immediate

aim

is to

set forth

 

277

027%4165/83 $3.00

 

Copyright

@ 1983 by Academic

Press. Inc.

 

AU rights of reproduction

in

any form

reserved.

  • 278 MC

GUIRE

AND

SCHIFFER

 

a testable, middle-range nacular architecture.

theory

to explain

variability

 

and change

in

ver-

 

Archaeologists

examine

the

end products

of design-particular

struc-

tures-that

 

can be characterized

by

formal

properties,

 

such as size, shape,

and construction

materials.

To explain

how structures

come

to have

spe-

cific designs-why

some

are large

and others

small, why

some are made

of

wood

and others

of stone,

why

some are internally

partitioned

and

others

not-we

must

examine

the

design

process.

In particular,

we

must identify

the general

causal factors

(and their

interrelationships)

 

that

influence

the decisions

leading

to

the

designs

for

specific

structures.

 

On

the

broadest

 

level,

of course,

availability

 

of materials

and

tech-

nology

constrain

architectural

designs.

Although

they

may have desired

it,

the builders

of Pueblo

Bonito

in Chaco

Canyon

could

not

have used

marble;

nor, with their technology,

could

the same Bonitians

have erected

a pueblo

of

50 stories.

However,

these

types of constraints,

which

put

generous

limits

on designs, furnish

relatively

few insights

into

the causes

of variability

between

societies,

and contribute

little

to explaining

differ-

ences or changes in the vernacular

architecture

within

societies (cf. Rapo-

port

1980a, 1980b). Given

the wide

limits

set by technology

and available

materials,

investigators

must

pay special

attention

to

the

social process

of design that determines

where-within

the limits-choices

actually fall.

The social process

is likely

to narrow

the options

considerably,

and,

on

occasion,

it

can

alter

the existing

constraints

by inventing

new technol-

ogies or

by securing

formerly

“unavailable”

materials.

Let

us now

begin

to build

a social theory

of design.

 
 

A SOCIAL

THEORY

OF DESIGN

 
 

Architectural

design

is a process

whereby

social groups

make choices

concerning

 

several

recurrent

sets

of

activities

(Rapoport

1980a:287,

198Ob:7). We focus

on

the

activity

sets of production,

use, and

mainte-

nunce

of the

built

environment.

With respect

set, people

attempt

to maximize

certain

goals.

Because

to each activity the activity

sets are

inter-

dependent,

 

it

is impossible

in

the

design

process

to

maximize

all

goals

simultaneously.

Moreover,

maximization

of

one goal

is usually

achieved

at

the expense

of others.

Thus,

the design process

 

can

be

viewed

as a

series of compromises

between

goals,

the result

of which

is necessarily

the achievement

of

some

goals at

less than

a maximum

level.

 

Factors relating

to a society’s

social structure

and adaptation

determine

the specific

content

and weighting

of the goals. Social differentiation

 

and

social inequality

affect

both

the utilitarian

and symbolic

requirements

of

 

THEORY

OF

ARCHITECTURAL

DESIGN

 

279

structures.

Residential

mobility

has a direct

effect

on utilitarian

require-

ments,

such as anticipated

uselife,

and constrains

the investments

 

that

can

be made

in architectural

symbolism.

In

the context

of these

other

factors,

architecture

must remain

responsive

to the developmental

cycles

of households

and institutions.

 
 

Design

is

a

social

process

because

compromises

between

goals are

effected

within

and between

social groups.

As societies

become

more

differentiated,

these

activity

sets and

their

attendant

goals become

in-

creasingly

associated

with

different

social units.

Among

societies

with

limited

differentiation,

such

as the

Navajo

(Jett and Spencer

1981: 17) and

the Tarahumara

(Pennington

1963:223,

226),

the family

that

will

use and

maintain

a structure

often

also builds

it, sometimes

with

such

assistance from

other

families.

 

In

more

differentiated

societies,

as the Samoans

(Goldman

1970:255),

some families

or task groups

specialize

in construc-

tion,

while

others-the

 

occupants-use

and maintain

the structures.

Fi-

nally,

in the most differentiated

societies,

such

as our

 

own,

separate task

groups

have arisen

that specialize

in design,

construction,

maintenance,

and even demolition

of structures.

In

this

case,

there

is often

appreciable

social distance

between

those

who

made and maintain

the structure

and

its users.

 
 

The separation

of these activities

 

among different

groups

increases

the

potential

for

conflict

in

the

design

process.

Even

in

the

simplest

case

where

a single family

or institution

 

participates

in all

three

activity

sets,

individuals

in the

group

may have divergent

views as to the requirements

for

use

and

on

the

compromises

to

be made

between

the goals

of

use,

production,

 

and maintenance.

Increased

differentiation

 

associates

each

activity

set with

a different

social collective,

each acting

in

its own

best

interest

to maximize

goals.

Since the goals of production,

use,

and main-

tenance

cannot

all be maximized

with

respect

to a particular

structure,

the potential

for

conflict

grows,

and

may

be exacerbated

by imperfect

communication between social units.

 
 

In the compromises

reached

during

the design process,

goals relating

to

use

are

usually

accorded

a high

priority

(Allsopp

1977:81-95).

 

Use

goals establish a series of requirements

with attendant

ranges within

which

the design

must

fall.

However,

where,

precisely,

the design

falls within

a

particular

range (and,

to

a certain

extent,

the definition

of the

range

itself)

depends

on the compromise

(or trade-off)

between

production

and main-

tenance .

 
 

Let

us now consider

 

some of

the causal

factors

that

shape the design

process

by examining

in

more

detail

the basic

goals

of

use,

production,

and maintenance.

We will

be stressing

the trade-offs

between production

and maintenance goals.

  • 280 MC

GUIRE

 

AND

SCHIFFER

Goals

of

Use

 

The goals

of

use can be partitioned

into utilitarian

and symbolic

func-

tions

(cf. McGuire

1981; Rathje and Schiffer

1982:67;

Saile

1977). This

dichotomy

 

is, admittedly,

an oversimplification,

but

it is nonetheless

a

useful

one

for

present

purposes.

The basic utilitarian

functions

of archi-

tecture

 

are

to

(1) mediate

between

people

(and

some of

their

artifacts)

and

the

natural

environment

and (2) delineate

space for the performance

of activities

by various

social units.

To carry

out these general

functions,

built

environments

must

meet

certain

requirements,

which

can

be enu-

merated

in terms

of specific

material

factors,

stemming

in part

from

the

nature

and diversity

of the activities

and social groups

that

use the space.

For example,

 

the diversity,

spatial extent,

and season of performance

of

activities

as

well

as the

size

of

the

social unit

lead

to minimum

floor-

space needs

(cf. Powell

1980; Schiffer

1973). Temporal

patterns

of struc-

ture

use along

with

diurnal

and seasonal temperature

variations

give rise

to specifications

for

building

materials

having

certain

insulation

values

(Evans

 

1980).

It

should

be recalled

that

the specific

material

factors

and

use goals

are not

always

realized

because

of raw

material

and technolog-

ical constraints

and compromises

in the

design

process.

From

 

the standpoint

of the archaeologist,

the delineation

of

use goals

and material

factors

in highly

specific,

quantitative

terms

may be difficult

given

our

present

knowledge.

A

few

examples

underscore

these

prob-

lems.

 

Although

a host

of studies

have begun

to establish

correlations

between

the

size

of

social

units

and amounts

of dwelling

space (for examples

of

such studies,

 

see Hassan

1981), we cannot

yet reliably

specify

the

min-

imum

space requirements

for

any particular

prehistoric

case. Clearly,

a

great

deal more

work

is needed

to discern

the influence

of activity

pat-

terns,

seasonality,

and other

factors

on the floor-space

requirements

of

dwellings

(cf. Schiffer

1972). Until

these studies are carried

out,

specific

formulations

of floor-space

requirements

must be regarded

as highly

ten-

tative

.

 

Even a mundane

requirement,

such as the maximum

load that

a wooden

roof

must bear, cannot

be translated

into quantitative

terms for

many

of

the construction

 

materials

used in vernacular

architecture.

Although

there

is a vast

body

of literature

on the mechanical

and decay-resistance

prop-

erties

of

lumber,

trees that

have little

or no commercial

value today

are

not

well

represented

(e.g.,

Bodig

and Jayne

1982). Thus,

if

one

wishes

to specify and compare

behavioral

requirements

for prehistoric

construc-

tion materials,

such as juniper,

 

pinyon,

mesquite,

and ironwood

in

the

 

THEORY

OF

ARCHITECTURAL

DESIGN

281

American

Southwest,

it may

be necessary

to carry

out new experiments.

In

the

absence

of

such data,

arguments

(such

as our

own

below)

neces-

sarily

rest

on

a very

insecure

footing.

 

Symbolic

functions

and requirements

of architecture,

 

which

facilitate

the workings

of ideology

and social structure,

are

not

as concretely

de-

finable

as utilitarian

functions.

In discussing

symbolic

functions,

we

are

not concerned

with

how a structure

becomes

imbued

with

meaning,

but

rather

with

how symbolic

requirements

enter

into

the design process

and

influence

the physical

form of architecture.

Buildings

with predominantly

 

utilitarian

functions,

such as Navajo

hogans,

can acquire

extensive

sym-

bolic content (Kent 1980). The Navajo hogan, however,

 

remains

a struc-

ture whose design-expediently

constructed

from earth,

stone,

or wood-

 

is mainly

utilitarian,

with

little

structural

investment

in symbolic

function,

despite its heavy symbolic loading (Jett and Spencer 198151-105).

 
 

Symbolic

functions

do under

certain circumstances

lead individuals

and

social groups

to make

investments

in architecture

beyond

or

in

spite

of

a building’s

utilitarian

requirements.

These investments

 

take

the form

of

decoration,

use of rare or expensive

materials,

 

building

on a grand

scale,

employing

a certain

shape (such

as the

cruciform

plan

of a basilica),

and

use of particular

 

construction

techniques.

It

is obvious

that

societies

vary greatly

in how

much

is invested

in such symbolism,

 

and the question

is why.

 
 

We propose

that

the

structural

investment

in symbolic

functions

 

in-

creases in response

to greater

social differentiation.

In societies

having

more groups

and more social distinctions,

there

 

is a need to communicate

 

ever more information

materially

(cf.

Wobst

1977; Hodder

1979,

1982;

Rathje

and Schiffer

1982). Furthermore,

as social units

 

become

increas-

ingly

specialized,

artifacts

with

high

symbolic

content-especially

 

built

environments

associated

with

religious

institutions-are

 

needed

to

help

integrate

a society’s

disparate

parts

(Rathje

and Schiffer

1982).

If

the

function

 

of

an institution

is primarily

ideological

or social,

then

its

in-

vestment

in architectural

symbolism

should

be relatively

greater

than

if

its functions

are primarily

economic

or technological.

 

Although

banks

and insurance

companies

in our

own

society

have mainly economic

func-

tions,

symbolic

elaboration

of architecture

is needed

to project

to poten-

tial clients

and stockholders

a corporate

image of “success”:

reliability,

wealth,

and permanence

(Duffy

1980). Indeed,

as Rathje

 

(1982) has noted,

even after

an institution’s

economic

base has eroded,

it

may

still

make

seemingly

irrational

investments

in material

symbols.

In

highly

differen-

tiated

societies,

few structures

are built

totally

 

without

symbolic

invest-

ment;

but

some have

more than

others,

depending

on their functions.

 
  • 282 MC

GUIRE

AND

SCHIFFER

 

Clearly,

with

 

respect

to

vernacular

architecture,

a change

in goals

or

in the

use requirements

of

or institution

will

lead

to

changes

in the

meaning,

symbolic

a social unit investment,

and

the

design

of architecture.

 

Goals

of Production

and Maintenance

 

The major

goals of production

and maintenance

are set forth

is to

with com-

parative

ease.

A primary

goal of production

 

minimize

the

cost

of

the manufacture

process

 

(Wilson

1971:26),

whereas

the principal

goal

of

maintenance

is

to

minimize

the

cost

of keeping

a structure

functional

during

its uselife

(Keiser

197856).

For

both

manufacture

and mainte-

nance,

cost

is defined

in terms

of energy

expended,

value of materials,

and expertise.

In most

real situations,

the goals of manufacture

and main-

tenance

come

into

direct

opposition.

Usually,

low

maintenance

cost

is

achieved

by greater

manufacture

cost,

and

low

manufacture

cost tends

to inflate

the

cost

of

maintenance

(Keiser

197856).

Because

of

this

re-

ciprocal

relationship,

 

compromises-sometimes

quite

painful

ones-

characterize

the design

process.

 

The

relative weighting given

to production

and maintenance

costs

in

design

is determined

by factors

pertaining

to

a society’s

social structure

and basic adaptation.

 

The

main

factor

from

social

structure

is social

inequality,

 

which

is

conveniently

viewed as relative

access to resources

(cf. McGuire

1983).

Access to resources determines

the ability

of

a social

unit

or institution

to mobilize

materials

and labor

for production.

For example,

the poor-

 

who

live

on a day-to-day

basis-have

little

call

on resources

and

can

establish

no foundation

for sizable investments

in production.

As a result,

an unhappy

compromise

 

is often

struck

between

production

and main-

tenance,

with

the goals of maintenance-and

 

sometimes

use-sacrificed.

 

Not

only can wealthier

groups

afford

to make

more

favorable

trade-offs

between

production

and maintenance

costs,

but they

rel-

atively

greater

amounts

to fulfill

the symbolic

can also invest use.

goals of

Increased

social inequality

has three

predictable

effects

on architec-

tural design

in a society:

(1) relatively

higher investments

by elite persons

and wealthy

institutions

 

in

the

symbolic

component

of architecture,

(2)

more

variability

in

the

production

costs

of architecture,

and

 

(3) more

advantageous

trade-offs

between

production

and maintenance

costs for

the structures

of

the

elite

and of wealthy

institutions.

for

 

These consequences

of

social inequality

can be expected

a number

of reasons.

At

the most

basic level,

greater

inequality

means that

the elite

have

at their

disposal

more

of

a society’s

total

production

to invest

in

structures

beyond

minimum

utilitarian

requirements.

As such,

the

elite

 

THEORY

OF

ARCHITECTURAL

DESIGN

283

can define

and reinforce

architecturally

their

dominant

positions

in

the

social

system

(Webster

1976:816).

Such displays

may take

the form

of

“over-”

building,

either in terms of scale or substantialness.

In this

manner,

large groups

of people

become

dependent

on the

elites for

their

livelihood,

 

which

helps

to maintain

the

elite’s

preeminent

position

(cf.

Schneider

1974). Conversely,

for

the mass

of population

that is relatively

deprived

of resources,

structures

may

meet

utilitarian

goals only minimally.

 

Fi-

nally, as Netting

(1982) establishes

cross-culturally,

wealthier households

tend

to

have

more

members;

thus,

they

would

need to construct

larger

dwellings

(see also Whiting

 

and Ayres

1968:122-123).

In

many

historic

societies

and those

known

archaeologically,

the

greater

investment

in

manufacture,

with

attendant

reduction

in maintenance

costs and

exten-

sion

of

uselife,

is reflected

in

what

often

survives

in good condition

 

for

investigators

to

study:

structures

of

elite

social

units

and wealthy

insti-

tutions.

Clearly,

the

distribution

of

wealth

in a society-social

 

in-

equality-intluences

both

the

mix

of utilitarian

and symbolic

functions

of

architecture

as well

as the trade-offs

between

manufacture

and mainte-

nance costs.

 

As social

inequality

changes,

so

too

should

a society’s

architectural

 

designs.

However,

certain

lag effects,

depending

on

the

uselives

of

a

society’s structures,

may

also occur.

 

For example,

a case can

be made

that

social inequality

in England

is somewhat

less today

than

that

indi-

cated

by

variations

in domestic

structures,

many

of which

were

built

centuries

ago

by

families

that

are

no

longer

affluent.

Indeed,

some

of

these

mansions

are

now

being

converted

into

museums.

Because

elite

structures

tend

to

have

longer

uselives,

the fortunes

of

the

elite relative

to

the

rest

of

society

may decline

without

being immediately reflected

in

architecture.

In an archaeological

time frame,

where rates of construction

for different

dwelling

types

through

time

can be ascertained

and where

changes

in function

can

often

be determined

using

the evidence

of

re-

modeling,

these lags should

be detectable.

 

The main

factors

of

a society’s

adaptation

that

influence

the relative

weighting

of production

and maintenance

costs are residential

mobility

(on household

and community

levels) and settlement

longevity.

These

in

part determine

 

a structure’s

 

anticipated

uselife.

Anticipated

uselife

is the

critical

variable

that

links

social

and adaptive

factors

in

the

decision-

making process.

The longer

the expected

uselife,

the

more

benefits

ob-

tained

from

a greater

investment

in production.

 

The interaction

of uselife

with

other

factors

that affect

design

is com-

plex,

but some relationships

 

can be posited.

We propose

that

the antici-

pated

longevity

of a settlement,

as determined

by a society’s

basic adap-

tation,

is the

most

important

 

causal influence

on the minimum

acceptable

uselife of structures.

When societies

are highly

mobile,

as is the

case with

most hunter-gatherers,

settlement

longevity

is generally

quite

short.

As

  • 284 MC

GUIRE

AND

SCHIFFER

one might

expect,

structures

are expediently

constructed

and ephemeral

(Sahlins

1972:3).

Indeed,

several cross-cultural

studies have shown

a re-

lationship

between

floor

 

plan,

substantialness

of dwellings,

and degree

of

residential

mobility

(Robbins

1966; Whiting

and Ayres

1968; cf. Flannery

1972; Rafferty

n.d.).

Generally,

the more

nomadic

the society,

the more

likely

it

is to

construct

domed

dwellings

with

round

or

oval

floor

plans.

Such structures

tend

to

be less substantial

and

have

less stringent

re-

quirements

for building

materials

than the rectangular

structures

that usu-

ally typify

long-lived

settlements.

As a result,

the uselives

of

such struc-

tures are generally

short.

 

The Znjluence

of Cultural

and Adaptive

Factors

on Production,

Use,

 

and Maintenance

 

These

relationships

between

settlement

mobility

and uselife

of dwell-

ings make

good

sense

in

terms

of trade-offs

between

production,

use,

and maintenance.

Highly

mobile

groups,

which

tend

toward

very

low

social inequality,

have

little

need

to express

social inequality

or social

differentiation

architecturally.

Nor

do these

societies

have

to maintain

structures

for

long

periods

of

time.

Thus,

factors

relating

to

the

basic

adaptation

and social

structure

of mobile

groups

favor

expedient

con-

struction

of dwellings.

The ephemeral

nature

of housing

for

most archae-

ologically documented

hunter-gatherers

is,

of course,

well

known.

 

A consideration

of domed

vs rectangular

architecture

 

permits

us

to

examine

some

of these

relationships

more

closely

and allows

us

to

set

the

stage for

the pithouse

to pueblo

example.

 

Domes

fit the utilitarian

housing

requirements

of mobile

groups

almost

perfectly.

Because

of

a greater

ratio

of volume

to surface

area,

domed

dwellings

require

less material

to construct

(Swanson

1981:vii-viii),

 

and

they

may

be made

of flexible,

lightweight

materials of irregular

shape

(Whiting

and

Ayres

1968: 124-125).

In addition,

hemispheric

 

structures

have a lower

wind resistance

than rectangular

buildings

(Keiser

1978:21),

and they

heat

up and

cool

down

more slowly

than rectangular

structures

of equivalent

size (Evans

198050).

These

characteristics

 

permit

people

who

build

domes

to achieve

their

basic utilitarian

goals of shelter

at low

cost.

These advantages

seem to imply

that domed buildings

are a universally

optimal

form

of structure;

indeed,

precisely

that

claim

was made

by

the

counterculture

builders

of

the

late

1960s

in

our

own

society

(Kahn

1978:200-201).

However,

domes

have

a number

of disadvantages

that

make

for bad compromises

 

between

production,

use, and maintenance

under

other

conditions.

If

people

anticipate

using

a structure

for

long

 

THEORY

OF

ARCHITECTURAL

DESIGN

 

285

periods

of time,

especially

a full generation

or more,

then

they will

benefit

from low maintenance

costs. Domed

 

structures

tend

to have higher

main-

tenance

costs because they are less substantial

and because they

are often

constructed

of perishable

materials

(Whiting

and Ayres

1968: 121).

It

is

difficult

to build

a dome

out

of stone,

adobe,

or brick,

all materials

that

require

less maintenance

than

the organic

items

customarily

employed

for making

domes

in societies

lacking

access to aluminum

and fiberglass

(Keiser

1978:56).

Although

domes

provide

more volume

per

unit

of

sur-

face area,

not

all of this

space is available

for activities

because

of a lack

of headroom

around

the margins

of

the structure

(Kahn

1978:200-201).

 

Moreover,

as Kahn

(1978:200-201)

 

also points

out,

domes

are difficult

to subdivide,

and this

is a distinct

liability

for less mobile

groups.

Internal

partitioning

of structures

facilitates

storage (Hunter-Anderson

1977), which

is more

important

 

to settled

peoples,

and

it provides

better

privacy

for

the building’s

Privacy,

defined

as control

of unwanted

social

interaction,

occupants. is a culturally

variable

 

concept;

nevertheless,

privacy

is rec-

ognized

and valued

to some

degree

in all

societies

(Rapoport

1976). Mo-

bile populations,

 

if

we

may

generalize

from

the Navajo

(Kluckhohn

and

Leighton

1962:91),

obtain

privacy

in

part

through

economic

and social

activities

that

often

take individuals

away

from

home

for

long periods.

In sedentary

settlements,

which

usually

have larger populations

and per-

form

most

activities

within

a day’s

radius

of

home,

privacy

may

be

achieved

by dividing

structures

internally.

 

Still other

reasons,

pertaining

to social organization,

can

be found

for

why domes

are not always

good

solutions

to the housing

problem.

House-

holds

in

all societies

go through

developmental

 

cycles:

an individual

or

couple

founds

a new household,

which

grows

by

the

addition

of

other

members

such as children,

relatives, friends,

and servants (Wilk and Rathje

1982). Even though

nonresidential

social groups,

 

like

task units

or

reli-

gious societies,

do not

always

undergo

developmental

cycles,

they often

experience

temporal

variation

in size (cf. Rathje

 

and Schiffer

1982). When

the composition

 

or

size of

social units

changes,

their

activities

(and thus

requirements

for

space) often

also change.

Architectural

design must

re-

flect

and adapt

to these variations

 

in

social

units

and

their

architectural

needs.

Hemispherical

structures

cannot

respond

well

to these changes,

whereas

rectangular

buildings

easily

accommodate

expansions

and

ad-

ditions.

This particular

deficiency

of domes

does not decrease

their use-

fulness

for populations

that

occupy

 

structures

for

only

a small fraction

of a household’s

developmental

cycle.

In

such

cases of

high

mobility,

growth