Top Banner
Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007
57

Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Mar 26, 2015

Download

Documents

Antonio Moran
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Can History Become a Real Science?

Peter TurchinUniversity of Connecticut

Talk presented at Santa Fe, March 2007

Page 2: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Main Points of the Talk• Most historians, philosophers, and

the lay public believe that there are no general laws of history

• I argue that the presence of strong empirical regularites implies the operation of general laws

• These laws can be discovered– there are much greater amounts of

quantitative data on historical processes than might be expected• but data sets are short and noisy

Page 3: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

The Focus of the Talk is

• not on past accomplishments– too early for that!

• but on future directions– what I intend to work on during the next

~5 years

Page 4: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 5: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

No General Laws of History?• Historical processes are too complex and

too different from physical or biological ones (Karl Popper)

• Any explanation of the course of events is specific to there and then (the great majority of historians)

• "There are no general laws in history, apart from those imagined by their proponents"

• "History is just one damn thing after another"

Page 6: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 7: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Ecosystems vs. Social Systems• Both are very complex and

heterogenous• Organisms have a kind of free will

– Insects, for example, are even less predictable than people

• At the micro level, ecosystems are a complete "mess"

• Yet, very clear patterns emerge at the macro level, such as population cycles– and there are laws of nature underlying

these patterns

Page 8: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Year

1950 1960 1970 1980 1990

Larc

h B

udm

oth

dens

ity

0.01

0.1

1

10

100

1000

Page 9: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Narrowing the focus: cliodynamics

• Large human collectives (≥105 ind)• Long time scales:

– a time step ≈ a human generation (20-30 y)

– dynamics on multi-decadal and centennial scales

• A key role for mathematical models• Quantitative variables, time-series data• Patterns at a macro scale, but

mechanisms at the individual level(There are other promising directions:

social evolution, micro-scale ABS, etc)

Page 10: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Cliodynamics vs. Cliometrics• Cliodynamics: from Clio (the muse of

history) and dynamics (the study of temporally varying processes)– an explicit math component (models)

• Cliometrics: in general, quantification in history– statistical, not mechanism-oriented;

lacks explicit theory-building approaches

• Synergism between the two approaches

Page 11: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Strong empirical patterns I• Secular cycles: second-order dynamics• The demographic-structural theory: a

rapidly maturing theoretical framework for explaining secular cycles– verbal propositions translated into a suite

of mathematical models– model predictions tested empirically for a

variety of agrarian states• strong effect of population pressure on real

wages (Malthusian mechanism)• strong effect of sociopolitical instability on

population growth

Page 12: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

England: the effect of population pressure on real wages

Year

1200 1300 1400 1500 1600 1700 1800

Var

iabl

es, l

og-s

cale

, arb

itrar

y co

nst.

"Misery index" = inverse real wagePopulation pressure = N/K

Page 13: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Instability Index (log-transformed)

-0.2 -0.1 0.0 0.1 0.2

Com

poun

d an

nual

gro

wth

rat

e

0.0

0.2

0.4

0.6

0.8

England: 1540-1870. Demographic data from Wrigley et al 1997 Instability data from quantification of narrative sources

Page 14: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Strong empirical patterns II:Religious Conversion

• Dynamics of many cases are well described by the logistic growth model

• Conversion to Islam– Iran– Spain

• Christianity• The Church of Latter-Day Saints

(Mormonism)- see Turchin 2003. Hist. Dynamics. Ch. 6

Page 15: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Iran: Bulliet (1979) Conversion to Islam

Century CE

7 8 9 10

Pro

port

ion

conv

erte

d

0

1

datamodel

Page 16: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Spain: Bulliet (1979)

Century C.E.

7 8 9 10 11

Pro

port

ion

conv

erte

d

0

1

datamodel

Page 17: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Mormonism: Stark (1984) The rise of a new world faith

1840 1860 1880 1900 1920 1940 1960 1980

Pro

port

ion

of W

orld

Pop

ulat

ion

Con

vert

ed

0.0000

0.0002

0.0004

0.0006

0.0008

0.0010datamodel

Page 18: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Strong empirical patterns III: Spatial distribution of

"imperiogenesis"

• Database: largest territorial polities– excluding modern sea-based empires

• Source: Taagepera, supplemented• Cut-off point: area ≥ 1 Mm2 at peak• More than 60 such polities are known

– only 1 (Inca) outside Afroeurasia

Page 19: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

M

Egypt

Axum

FatimAlmorav

Almohad

Mali

Mam

Hsnu

Juan

Turk

UigTufan

Khazar

Hsi

Khorezm

Kara-Kh

Mongol

GoldenH

ChagataiTimur

ShangHanTang

Liang

Liao

Sung

Jur

Ming

ManchuRom

HunsFrank Kiev

Lith-Pol

Osman

Russia

Srivi

Khmer

Maur

Kushan

GuptaHarsha

Delhi

Mughal

Mar

AssyrMed

AchSas

SeleParth

CaliphSelj

Sam BuyGhazn

AyyIl-Kh

Byz

Page 20: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 21: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Largest territorial polities tend to arise at interfaces between settled and nomadic societies

• Not a strict "law", but rather a statistical correlation

• Several "hotspots" of imperiogenesis and upsweeps in max. territorial size– Mesopotamia and Iran– Northern India– Northern China

Page 22: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Unification Period Ethnicity From Capital

Shang 1766–1122 BCE

?? NC (Huang He) Anyang (Huang-He)

W. Zhou 1122 –771 BCE Frontier Han(“Western Barbarians”)

NW (Wei River Valley)

Loyang (Huang He)

Qin 221–206 BCE Frontier Han NW (Wei River Valley)

Xianyang(Wei)

Han 202 BCE–220 Han NW (looks like their base was at the confluence of Wei and Huang)

Chang’an(Wei)

N. Wei(partial, N)

386–534 To-ba (Turkic) NW Loyang (Huang He)

Sui 581–618 Han NW (Wei River Valley)

Chang’an(Wei)

Tang 618–907 Han (ruling family of Turkish descent)

NW (Wei RV?) Chang’an(Wei)

Page 23: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Liao(partial, N)

907–1125 Khitan (Altaic?)

NE Beijing

N. Sung(partial, w/o N)

960–1127 Han NC (from lower Huang He area around Kaifeng)

Kaifeng(Huang He)

Jin(partial, N)

1115–1234 Jurchen (Tungus)

NE Beijing

Yuan 1206–1368 Mongol NW (Mongolia) Beijing

Ming 1368–1644 Han CS (from Nanjing area): the only unification not from N

Beijing

Qing 1644–1911 Manchu (Tungus)

NE Beijing

Communist 1949– Han NW (Long March to Wei River Valley; unification from there)

Beijing

Page 24: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

The East Asian Imperiogenesis Hotspot: Empirical Patterns

• 14 unifications of China from the Shang to Communist eras (some partial)– (E.N. Anderson, supplemented)

• Summary:– 8 unifications from NW (usually, Wei RV)– 3 unifications from NE (Liao, Manchuria)– 2 unifications from NC (Huang He)– 1 unification from SC (Nanjing)

Page 25: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

The broad context:The puzzle of human

ultrasociality• Evolution of cooperation in small

groups (~102 ind) by group selection is essentially understood – D.S. Wilson, Boyd, Richerson, Bowles

• Asabiya (Ibn Khaldun): capacity for collective action

• But how did huge groups of 106−108 cooperating individuals arise?

Page 26: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

The "Mirror Empires" Model

• A steppe frontier between settled agriculturalists and nomadic pastoralists

• Starting point: small-scale polities on both sides of the frontier

• Pastoralists enjoy preponderance of military power; need the products of agriculture

Page 27: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 28: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 29: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 30: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 31: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Outcome• An agrarian empire and a nomadic

imperial confederation arise simultaneously in a mirror fashion

• The process occurs in a series of steps of increasing territorial size and social complexity

• A positive feedback loop (self-feeding process)

• Runaway territorial growth is eventually stopped by space or logistic limits

Page 32: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Two Kinds of Sciences(Randall Collins, The Sociology of

Philosophies)"Rapid Discovery" "Traditional"

Rapid rate of knowledge production

Slow rate of knowledge production

Consensus Dissensus

Priority disputes; simultaneous discovery frequent

Disputes focus on clashing ideologies

Nobody reads the founders

The founders are constantly re-examined and re-appraised

Page 33: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Can cliodynamics become a "rapid discovery science"?

• In the end, this is an empirical issue: "the proof is in the pudding"

• We have to generate a constant flow of new results

• We need to propose and defend candidates for general laws

• So what about the data?

Page 34: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Sources of data

• Archaeological• Skeletal material• Coin hoards• Quantification of narrative sources• and many, many other

Page 35: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Novgorod the Great

Page 36: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 37: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 38: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Novgorod: Time distribution of birchbark documents

1000 1100 1200 1300 1400 1500

Doc

umen

ts p

er d

ecad

e

0

10

20

30

40

50

60

Page 39: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Nerev

1000 1100 1200 1300 1400 1500

Nor

mal

ized

dat

a

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

shoeslocksclothamber

Page 40: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Stature as a proxy for population density

• Abundance of data (106 skeletons)• Human height is a very sensitive

indicator of nutrition conditions – a proxy for population pressure

• But temporal resolution is poor– Radiocarbon dating errors are ~ 50 y– However, given the abundance of data,

it should be possible to use statistical methods for error reduction

Page 41: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Average height: skeletal material from Europe (Koepke et al 2005)

Birth Century

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Hei

ght i

n cm

(in

vers

e sc

ale)

168

169

170

171

172

173

Principate

Dominate

Carolingian

Early Medieval

LateMedieval

EarlyModern

Page 42: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Skeletons, cont.

• Can be used to score the intensity of interpersonal violence, and thus, indirectly, sociopolitical instability

• Example: the study of Tim Kohler et al in the American Southwest

Page 43: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Figure 4. Graph of standardized, smoothed population (N, black) superimposed on smoothed warfare frequency (W, red). (Kohler et al. 2006)

Page 44: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.
Page 45: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Coin hoards• Abundant in many historical eras;

datable• Frequency of hoards (per decade)

reflects conditions of internal disorder:– people bury hoards in times of danger– most emergency hoards are recovered,

except when the owner is unable to do so• Caveat:

– hoard incidence reflects not only internal disorder, but also catastrophic external invasions

Page 46: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Year, BCE

-200 -150 -100 -50 0

Hoa

rds

per

deca

de

0

10

20

30

Coin Hoards: Republican Rome, 230-0 BCE (Michael Crawford)

Page 47: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Year, BCE

-200 -150 -100 -50 0

Inst

ab

ility

In

de

x

0

2

4

6

8

10

Instability in Republican Rome, from narrative sources

Page 48: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Year

-220 -200 -180 -160 -140 -120 -100 -80 -60 -40 -20 0

Coi

n ho

ards

0

10

20

30

Inst

abili

ty I

ndex

0

2

4

6

8

10HoardsInstability

Coin Hoards and the Instability Index

Page 49: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Bohemia, Moravia, and Silesia

Year

1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850

Hoa

rds

per

deca

de

0

50

100

150

200Hussite Wars (1420-36)

Thirty-Year War (1618-48)

Page 50: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Hoards in NW Germany

1000 1100 1200 1300 1400 1500 1600 1700 1800

Hoa

rds

per

half-

cent

ury

0

10

20

30

40

50

W. WestfalenE. WestfalenPfalz/SaarNordrhein

Welf/Hohen-staufencivil wars

LateMedievalCrisis

Crisis of XVIIc

Page 51: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Quantifying sociopolitical instability from narrative

sources• Sociopolitical instability:

– state collapse, peasant uprisings, civil wars, and other instances of major internal disturbances

• Construct an annual index by noting which years had an instability event, and which years did not (either 0 or 1)

• A decadal index = the number of instability years per decade (varies between 0 & 10)

Page 52: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

years Description

1138-53 Anarchy (civil war between Stephen and Matilda/Henry)

1173-4 Uprising of Henry the Younger. Rebellion of several English earls.

1215-7 Civil war (Magna Carta)

1232 Revolt against papal collectors

1233 Richard Marshal rebelled and was murdered in Ireland

1263-7 Civil war: barons against the king

1315 Civil disorders during supremacy of Lancaster (1314-22)

1321-2 Civil war. Baron uprising in the western counties.

1326-7 Rebellion of Isabella and Mortimer. Edward II deposed, murdered in prison

1330 Edward III led the baronial opposition to Mortimer (hanged, 1330)

1381 Peasants' Revolt

1387-8 Insurrection of the “Lords Appelant”

1391 Coup d’etat of Richard II

1397-9 Events leading to the deposition of Richard II (1399).

1400-8 Glyn Dwr rebellion

1414 A Lollard plot against the king's life

1448-51 Domestic disorders

1450 Jack Cade's rebellion

Page 53: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

years Description

1455-6 The Wars of Roses: 1st phase

1460-5 The Wars of Roses: 2nd phase

1467-71 The Wars of Roses: 3rd phase

1483-5 The Wars of Roses: 4th phase

1495 Rebellion of Perkin Warbeck

1497 Insurrection in Cornwall

1536-7 Pilgrimage of Grace

1549 Kett’s rebellion

1554 Wyatt’s rebellion

1569 Rebellion of catholic lords of the North

1639-40 The Bishops’ Wars

1642-7 Civil War

1648-51 Second Civil War

1655 Penruddock rising in Salisbury

1660 Monk’s coup; restoration of James II

1666 Revolt of Scottish Covenanters

1679 Revolt of Scottish Covenanters

1685 Monmouth and Argyll rebellions

1687-92 Glorious Revoultion, with intervention by France

1715-6 Jacobite rebellion in Scotland

1745-6 Scottish rising (Jacobite pretender)

Page 54: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

England

1100 1200 1300 1400 1500 1600 1700 1800

Inst

abili

ty I

ndex

0

5

10

Events per decadeSmoothed, h =25 y

Page 55: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Rome: Sorokin's index of internal war

Year

-500 -400 -300 -200 -100 0

So

cio

po

litic

al i

nst

ab

ility

ind

ex

0

50

100

150

200

250

Page 56: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

"Wheels within Wheels"

• Two kinds of oscillations superimposed:

• Secular cycles– periods = 200−300 y, or ~10

generations (second-order dynamics)

• "Fathers-and-sons cycles"– periods = 40−60 y, or ~2 generations

(first-order dynamics)

Page 57: Can History Become a Real Science? Peter Turchin University of Connecticut Talk presented at Santa Fe, March 2007.

Main Points of the Talk• Most historians, philosophers, and

the lay public believe that there are no general laws of history

• I argue that the presence of strong empirical regularites implies the operation of general laws

• These laws can be discovered– there are much greater amounts of

quantitative data on historical processes than might be expected• but data sets are short and noisy