The Two Poverty Enlightenments: Historical Insights from Digitized Books Spanning Three Centuries

Word searches of Google's library of digitized books suggest that there have been two "Poverty Enlightenments" since 1700, one near the end of the 18th century and the second near the end of the 20th. The historical literature suggests that only the second came with a widespread belief that poverty could and should be eliminated. After the first Poverty Enlightenment, references to "poverty" (as a percentage of all words) were on a trend decline until 1960, after which there was a striking resurgence of interest, which came with rising attention to economics and more frequent references to both general and specific policies relevant to poverty. Developing countries also became more prominent in the literature. Both Enlightenments came with greater attention to human rights. The written record reflects the push-back against government intervention and the retreat from leftist economics and politics since the late 1970s. Although many debates from 200 years ago continue today, there is little sign that the modern revival of the classical 19th century views on the limitations of government has come with a revival of the complacency about poverty that was common early in that century.


Introduction
How has public attention to poverty changed over time? Has less poverty in the world as a whole deflected that attention or enhanced it? How has it reflected economic and social thought more broadly? Were some periods more enlightened about poverty than others? Prevailing historical narratives that might throw light on such questions have relied heavily on the careful human reading of texts. While this is understandable, it seems that historians might also exploit the digital age. 2 The technology of digitizing texts has opened up new options for systematically searching words and phrases, as a complement to more traditional historical research methods.
A new window has recently been opened on the prominence of ideas in the writings of their time. The Google Books Ngram Viewer (the Viewer hereafter) was introduced by Jean-Baptiste Michel et al. (2010). In this context, an "n-gram" is just a phrase, made up of words ("1grams"). Michel et al. created a corpus of n-grams (for n=1,5) from 5.2 million books (excluding periodicals). The Michel et al. corpus holds over 500 billion words (360 billion in English). By conventional human reading, it would take the working lives of 347 people to read the entire corpus at 300 words per minute. 3 The Viewer is fast; word searches and plots over hundreds of years take only a small fraction of a second. Naturally the total number of words in a given year has increased over time. Michel et al. estimate that there were about one million words in English in 2000, but barely half that number in 1900. (And most of that growth was in the latter half of the 20th century.) The Viewer"s counts are normalized by the total number of n-grams in the corpus for that year. I will refer to the normalized frequency as the "incidence." This paper uses the Viewer to study the evolution of attention to poverty and related topics in the literature. The vast majority of Google Books are not, of course, specifically about these subjects. What we learn from reading them is more about the public"s written attention to a topic, which need not correspond well with influence in more specific circles, such as academics or policy makers. (Nor does awareness, as reflected in language, imply action.) This is not a limitation for the present purpose, since Google Books is being used here as a linguistic window on public awareness. If certain time periods were critical in the history of popular thought on 2 Maxwell (2010) comments that historians seem to have been slow to take up the scope for analysis using digitized texts. 3 I assume that they would work eight hours a day for 250 days per year for 40 years.
3 poverty-such as the revolutionary period of the late 18th century, or around the late 19th century when scientific studies of poverty emerged in the social sciences-then we should see an upsurge of references to "poverty" in the general literature. Similarly, if (as some observers have argued) the late 20th century has seen depleted concern for poverty and inequality then we should see that in the Viewer. Figure 1 gives the Viewer's graph for "poverty" from 1700-2000. This is a moving average of the incidence of that word in the English corpus by year. Three observations can be made about Figure 1. First, there was a large growth in references to poverty in the latter half of the 18th century, peaking around 1790. Toward the end of the Age of Enlightenment, around the time of the French and American Revolutions, came what I will term a "Poverty Enlightenment." Second, there was a clear trend decline in references to "poverty" (relative to the volume of all words) through the 19th and 20th centuries up to around 1960. Over these 170 years, poverty slowly faded from public attention. The third observation is that there was a significant awakening of that attention after 1960-a second Poverty Enlightenment. The initial sharp rise of interest peaked in the early 1970s, with a decline thereafter, followed by a resurgence starting in the mid-1980s, with the overall peak just a couple of years before the turn of the 21st century. The rest of this paper will try to understand Figure 1, drawing on both my reading of the historical literature and the Viewer. After discussing some of the data and methodological issues in using the Viewer, the paper discusses the history of thought on poverty and policies for fighting it, drawing on historical texts. This will provide some clues to understanding Figure 1,4 and also point to some questionable claims in past historical narratives. The paper then documents the time series of references to a number of n-grams related to poverty. The rise and fall of the various schools of economic thought is documented, as is the evolution of views on the role of the state and specific policies for fighting poverty. The set of graphs will also allow an assessment of the robustness of some of the paper"s earlier claims.

Using the NGram Viewer as a research tool
I began by identifying a series of seemingly relevant topics (which will unfold in the coming sections) and a list of key n-grams for each, with the same n-gram appearing more than once. The choice is not unconstrained. The same word or phrase can have different meanings at one date or changing meaning over time, and sometimes the word is so ambiguous as to be useless for the present purpose. "Poverty" appears to have had a relatively stable usage over time; when (say) Samuel Johnson writes about "poverty" in the mid-18th century he appears to be talking about essentially the same concept found in modern usage. Even so, one also finds some n-grams using the word "poverty" that have a different meaning, though they have low frequency. 4 There are other words, such as "poor," "misery," "rich" and "welfare" that are too ambiguous to be useful. "Inequality" is probably more ambiguous than "poverty" (there is, for example, the usage of "inequality" in mathematics), but is nonetheless too important to exclude.
While I have tried to select with care, I am sure that I have picked up some unintended or ambiguous usages. I will test whether my claims about trends are robust to acknowledging the use of other words for "poverty" at various times, such as "indigence," "destitution," "working class." (The latter term does not mean "poverty" but was often used to refer to poor people in the 19th century, notably in the literature on socialism.) Sometimes I chose to drop one of my selected n-grams for a given topic either because its frequency was so low relative to others in the series as to be of little interest, or it was evidently so synonymous with another n-gram that there was little new information (singular and plural 4 One example is the expression "poverty of ideas" (including at least one example from Dr Johnson"s writings). Even at its peak usage the incidence of the expression "poverty of ideas" was only 0.1% of the incidence of "poverty." Another example is the phrase "poverty of philosophy," which was the title of Karl Marx"s famous critique of Pierre-Joseph Proudhon"s Philosophy of Poverty, published in 1847. But, again, this is a very low-frequency n-gram. forms were common examples). 5 When there are multiple n-grams with essentially the same meaning I have chosen the most common one (though in some cases for which it seemed important, I give more than one). The Viewer is case sensitive; I have often checked this in preparing the graphs, and sometimes report both cases.
The database has digitized books back to 1500, although less than 10% were published before 1800 and it is clear that there is considerable noise in the data prior to 1800, especially when we turn to more specific aspects of poverty and policies for fighting it. This reflects the lower volume of books and lower quality of the optical character recognition (OCR). When the paper starts to take a more finely-grained look at poverty and policies it will focus on the period 1800-2000 and sometimes just the 20th century.
The scales of all graphs are cardinally comparable, each one giving the share of the ngram in the total number of n-grams for that year. However, the ranges vary between graphs. I have sometimes modified my initial groupings of n-grams to narrow the ranges of their graphs, and so make them more readable.
Naturally, there are year-to-year fluctuations in the word count, and there can be pronounced spikes, sometimes reflecting just a few books that mention "poverty" many times or a low overall volume of words that year. Spikes are more common prior to 1800. There are also lags in publication, such that references to a word may reflect its prominence one or even two years earlier. So some smoothing is called for. The Viewer allows alternative smoothing parameters defined by the number of years either side that one forms the moving average of normalized counts. 6 The graph in Figure 1 has been smoothed using a smoothing parameter of 4 (a moving average with a "bandwidth" that is four-years either side). Contrast this with the very erratic raw series here, and the much smoother graph here using a 20-year smoothing parameter, which makes the two Poverty Enlightenments stand out clearly, but loses some interesting details. I have chosen relatively low smoothing parameters in this paper, namely 0-5 (on a scale 0-50), depending on the appearance (setting a slightly higher value if it looked under-smoothed). 5 The Viewer only reports counts for n-grams that appear in at least 40 books (over all years). 6 For example, a smoothing of 2 means that the data shown for 1970 (say) will be an average of the raw count for 1970 plus 2 years either side, divided by 5. A smoothing parameter of "0" uses only the raw count for each year. The averaging is truncated at the edges of the graph, with fewer values averaged.

6
There is a degree of subjectivity in all such choices. Each graph in this paper gives the web link (embedded in its title) to the corresponding Viewer page, so readers can readily test the robustness of my claims.
One must consider some possible biases in the Viewer. Is it representative of the views of society at any date? There are two potential sources of bias to note. First, there is the question of whether the Michel et al. sample of 5.2 million books is representative of all books. Google Books now include 15 million digitized books drawn from 40 university libraries around the world, with supplementation from publishers. The Michel et al. sample was chosen according to the quality of both the OCR and the metadata (such as date and place of publication). The notes on the Viewer"s website claim that they did "a good job at filtering out books with low OCR quality scores," though they also acknowledge some errors (such as some pre-1950 references to the "internet"). No other selection criteria are mentioned by Michel et al. Given that Google Books has relied heavily on university libraries, it might be conjectured that the sample is more representative of the scholarly literature. However, popular books do seem to be well represented. The fact that the Google Books site is linked to retail sales options also suggests there will be an incentive to represent the popular literature.
A second possible source of bias is that books need not be representative of current thinking in the population. Books can only written by literate people (by definition). A literature written and read by relatively well-off people need not represent well the views of the population at large. Mass literacy emerged in Britain, the US and much of Western Europe around the turn of the 20th century (Vincent, 2000); around 1800, books could be produced by only about one half the population of England or France (say). And mass literacy is only recently emerging in the bulk (though still not all) of the developing world. With rising overall literacy rates, the literature of the late 20th century is almost certainly coming from a broader cross-section of society than that of the early 18th century. To the extent that this yields more references to poverty there will be a selection bias for inferences about overall public attention to poverty, including among groups in society that contribute little to the literature.
The spread of mass literacy must surely have impacted on the literature. While we don"t know the counterfactual, there is no sign in Figure 1 that the emergence of mass literacy in the developed world in the early 20th century led to a rise in awareness of poverty; the surge of 7 interest appears to have come much later. But possibly there are long lags. It is likely that the English-language literature of the late 20th century is more representative of the Englishspeaking world as a whole than that in 1800, which was far more focused on Britain, Western Europe and America.
There are other factors to be aware of in using the Viewer. It does not allow stratifications of the corpus beyond language, so we can only make limited conditional searches beyond language. (One can search books published in America, which are identified as "American English" in the Viewer window for language; the same holds for "British English.") One cannot identify how a given word count is distributed across books. And the new tool has little or no ability to directly measure linkage between ideas, such as when one set of ideas influences (positively or negatively) another. That requires more careful reading. And there are subtleties of usage and context that could probably never be detected by a machine.
While recognizing these limitations, the Viewer allows us to systematically quantify the evolution of public awareness of ideas, as revealed by the literature. The new tool allows us to trace the flow of ideas, and their relative influence. I will focus mainly on the English language, which has been the main language of the social sciences in the 20th century. However, given the prominence of the French literature on inequality (in particular), I will also make some use of the French language Google Books.

Historical perspectives on the two Poverty Enlightenments
What might account for the surge in attention to poverty in the last half of the 20th century in Figure 1? I can suggest two factors. The first is the sharp rise in concern over poverty in affluent societies that built up from the 1960s, notably in the US. In the wake of the civil rights movement (starting around 1955), the rediscovery of poverty in the midst of affluence in mid-20th century America was stimulated by important social commentaries, including John Kenneth Galbraith"s (1958) The Affluent Society and Michael Harrington"s (1962) The Other America (both best sellers at the time). These helped initiate a political response in the US, including new social programs targeting poor families, under the Johnson administration"s War on Poverty (Sundquist, 1968). References to both Galbraith"s The Affluent Society and Notice that there is a pronounced "bubble" in references to poverty, peaking around 1970 ( Figure 1). Consistently with this first factor, I conjecture that this bubble reflects the shock of re-discovering poverty in America. The success of Harrington"s book was a surprise at the time; the first print-run was 2,500 copies, but by the mid-1990s it had sold 1.3 million copies. In due course, the trajectory returned to normal. Of course, if one "smoothes out" this bubble sufficiently, then one sees a clear upward trend in references to poverty starting around 1950 through the end of the century, as can be seen here (using a smoothing parameter of 20 years).
The influence of these writings depended, of course, on their historical context.
Harrington and Galbraith clearly wrote their books at the right time to have the influence they had, given the combination of new affluence (for the majority) and optimism (including about policy) that prevailed in America in the 1950s and "60s. But as times changed, a backlash emerged, as has often happened. The counter-attack came in the 1980s; Charles Murray"s (1984) Losing Ground was an example in the written record. Welfare reforms followed in the US in the 1990s-reforms that were contentious among poverty scholars and advocates in the US. 30 years after declaring a "War on Poverty," America declared a "War on Welfare." 7 There are continuing debates today about poverty-its causes and the appropriate policy responses, throughout the world.
The second factor in the surge of attention to poverty in the late 20th century was the public"s increasing awareness of the existence of severe and widespread poverty in the developing world. Poverty in developing countries started to attract mainstream scholarly attention in the 1970s. 8 The World Bank"s (1990) World Development Report: Poverty was influential in development policy circles, and soon after a "world free of poverty" became the Bank"s overarching goal. A large body of empirical research on poverty followed in the 1990s.
The UNDP"s Human Development Reports began in 1990, and have argued for public action to promote basic health and education in developing countries. 7 The latter term was used by Katz (1987); also see Albelda et al. (1996). 8 Important contributions in the period 1970s and"80s included Dandekar and Rath (1971), Adelman and Morris (1973), Chenery et al. (1974), Lipton (1976), Fields (1980), Kakwani (1980), Sen (1981), Anand (1983), Bardhan (1984) and Kanbur (1987). While there has been much debate on the causes of poverty and on policy prescriptions, modern writings are invariably premised on a belief that poverty is something that can be greatly reduced with the right economic and social policies and, indeed, eliminated. By this view, poverty is in no small measure a public responsibility, and the government and the economy are to be judged (in part at least) by the progress that is made against poverty.
That view was not always held. 9 In his history of the idea of "distributive justice" Samuel Fleischacker (2004, p.7) argues that in pre-modern times, "the poor appeared to be a particularly vicious class of people, a class of people who deserved nothing." A seemingly common view among social commentators was that the pervasive poverty of the time was a more-or-less inevitable "fact of life." In the early 18th century, Robert Moss instructed the poor man "to rest contented with that state or condition in which it hath pleased God to rank him." 10 The French doctor and moralist Philippe Hecquet wrote in 1740 that "The poor of a State are like the shadows in a painting: they provide the necessary contrast." 11 To the extent that any effort were made to explain poverty it was seen as either "God"s will" or a purely private matter, stemming from bad personal behavior, such as laziness. Indeed, Bernard de Mandeville and (later) Joseph Townsend argued that hunger was a good thing, as it motivated poor people to work. 12 The late 18th century saw, as Fleischacker (2004, p.53) puts it, "a sea change in attitudes toward the poor." A new questioning of established social ranks was found among the emerging middle class, notably in France; in the 1780s, The Marriage of Figaro, a play by Pierre Baumarchais, had Parisian audiences taking side with the servants in laughing at the aristocracy.
(The play was written in 1778 but had been censored for many years. It is widely seen as a precursor to the French Revolution Novels started to document the lives of ordinary people.
Some of the new egalitarian spirit was spilling over the English Channel, though it met with still resistance. In 1806 Patrick Colquhoun, the founder of the police force in England, wrote that: "Poverty ... is a most necessary and indispensable ingredient in society, without which nations and communities could not exist in a state of civilisation. It is the lot of man-it is the source of wealth, since without poverty there would be no labour, and without labour there could be no 9 The following discussion will be largely confined to the period 1700-2000. There is an interesting discussion of "pre-modern" thought on poverty in Geremek (1997). 10 Quoted by Ashcraft (1996, p.26 riches, no refinement, no comfort, and no benefit to those who may be possessed of wealth." (Colquhoun, 1806, p.7).
The word "poverty" often appeared in both English and French literary works of the time.
Measured by the incidence of references to "poverty" in English-language Google Books, the peak of the First Poverty Enlightenment was 1787. There are 204 books listed as containing a reference to poverty in the database for 1787 used by the Viewer-representing 0.005% of all words that year, which was the highest incidence in a single year over 1700-2000. The 204 books containing wide-ranging references to poverty, and no single volume dominates, though the largest count is found in Samuel Johnson collected works (published after his death in 1784). Dr Johnson wrote (often amusingly) about the foibles and hypocrisies found in how people (often rich people) think and talk about poverty. While an awareness of poverty is certainly evident in these writings, it is by-and-large an accepting and unchallenging awareness. While the first Poverty Enlightenment saw far greater attention to poverty in the literature, it was not associated with widespread empathy for poor people as fellow citizens on the part of those writing books, and it was not yet widely accepted that there was any public responsibility for poverty.
Nor did the economics of the 18th and 19th centuries offer any serious challenge to the view that poverty was to be largely accepted. Workers were often equated with "the poor." Some economists saw poverty as an essential condition for economic development. No doubt it was agreed that rising real wage rates of the proletariat would reduce poverty, but it was argued that this would undermine wealth accumulation by reducing labor supply and (in the Mercantilist schema) making exports uncompetitive, and even corrupting the values of workers as they aspired to luxury goods. 13 Thomas Malthus famously saw ecological disaster ahead, with poverty and famine as the only check against rising population. It does not seem that even Adam Smith-widely seen at the time to be far more optimistic than Malthus about the scope for overall social progress-entertained much hope that the fruits of economic development would be equitably distributed: 14 13 Muller (1993) provides an interesting discussion of these and other views held in the late 18th century. The Mercantilists saw foreign trade as a zero-sum game, so anything that reduced a country"s trade balance would reduce its share of global wealth. 14 This is not to say, of course, that Smith cared little about poverty. Unlike some of his peers, Smith put weight on attainments of both the basic consumption needs of society as a whole in assessing national wealth, not just the luxuries available to the rich; see the discussion in Muller (1993, p.58). Also see Himmelfarb"s (1973) discussion of Smith"s views relative to others around the time, including Malthus.
"Whenever there is great prosperity, there is great inequality. For one very rich man, there must be at least five hundred poor, and the affluence of the few supposes the indigence of the many." (Smith, 1776, p.232).
Essentially the same view was held by the schools of economic thought more critical of capitalism, as emerged in the mid-19th century. Socialists also held that poverty and inequality were inevitable under capitalism, and could only be redressed by a socialist economy. The widespread acceptance of poverty in the 19th century, and even indifference to it, has been identified by Carl Landauer (1959) as one of the factors that led to the emergence of socialism.
The explicit acceptance of poverty among economists and non-economists alike some 200 years ago does not appear to be the product of a strongly relativist view of what "poverty" means, whereby standards for defining poverty rise with overall affluence. While Smith is often cited for his views on the social-inclusion role of a linen shirt in late eighteenth century Europe (Smith, 1776, Book 5, Chapter 2, Article 4), this merely suggests that he wanted a poverty line to be relevant to the society in which poverty is being estimated. The idea of a relative poverty line that is set as some constant proportion of current average income is a far more recent idea in Western Europe. Poverty appears to have been seen as absolute in the 19th century.
The classical economics that came into the ascendency in the 19th century, and with great influence on policy, held that a laissez-faire capitalist economy was essential for wealth generation. However, poverty and inequality appear to have been broadly accepted as its moreor-less inevitable byproduct. Malthus saw little scope for increasing the living standards of workers above subsistence. David Ricardo built on Smith"s (enormous) contribution, but did not share Smith"s optimism; like Malthus, Ricardo saw little scope for better living conditions for the working class-essentially equated with "the poor"-under capitalism, which (like Smith and Malthus) he strongly supported. Neither the scholarly nor popular views at the time saw capitalism as the route out of poverty. The industrial revolution brought technical innovation and new wealth, but it was not widely believed that industrial workers shared much in these gains. Engels (1845) and others described the poor health environments and harsh working conditions of English industrial cities in the mid-19th century and workers were not thought to have seen 12 much gain in their real wages. While these views appear to have been exaggerated or even wrong, 15 it appears that poverty was widely seen as a necessary condition for capitalism.
Nor did economics offer much support for direct interventions to help poor people. Adam Smith was something of an exception. He supported limited public subsidies to help cover tuition fees for the basic schooling of the "common people" (Book V, Chapter I, Article 2d). And while the utilitarianism that emerged around this time was explicitly indifferent to inequality of welfare and to individual rights, it did offer the germ of a theoretical justification for income redistribution, given diminishing marginal utility of income; for example, Smith was also supportive of progressive taxation in the form of property taxes (Book V, Chapter 2, Article 1).
The English "Poor Laws" were much debated around the turn of the 18th century. These had existed back to the 16th century, and provided a system of locally-implemented (Parishlevel) relief available to all. These "Old Poor Laws" had become contentious by the early 18th century. 16 Smith recommended reforms, notably to relax the settlement restriction, whereby relief was confined to local residents (a restriction on mobility). Other economists, including Malthus (1806, Chapters 5 and 6) and Ricardo (1817) argued for more radical reforms. Ricardo (1817, p.61) wrote that the impact of the Poor Laws is not, "as the legislature benevolently intended, to amend the condition of the poor, but to deteriorate the condition of both rich and the poor; instead of making the poor rich, they are calculated to make the rich poor; and whilst the present laws are in force, it is quite in the natural order of things that the fund for the maintenance of the poor should progressively increase until it has absorbed all the net revenue of the country." Adverse incentive effects (including early marriage and high fertility) have been a common argument against direct anti-poverty policies. The extent of these effects is unclear, and there are many potentially positive benefits that should also be considered, including in providing insurance and promoting economic development (as pointed out by Peter Solar, 1995, in the context of the Old Poor Laws). However, influenced by Malthus and Ricardo, significant reforms to the Poor Laws were implemented in the 1830s, notably though wider use of "workhouses," whereby recipients had to be domiciled in a workhouse, which was deliberately made uncomfortable as a form of "self-targeting." These were intended for the poorest, and for 15 See Landauer (1959) and the evidence compiled by Williamson (1985) suggesting that real wages had in fact risen appreciably between 1800 and 1850, due in large part to cheaper wage goods. 16 On the history of the English Poor Laws see Boyer (2002). the "deserving poor" in particular, not as a general remedy for pervasive poverty. The scope for redressing poverty and inequality through either laissez-faire capitalism or social policy was clearly seen to be strictly limited for the bulk of the 19th century.
One is struck by how much the modern policy debates have in common with those of the 19th century. Incentive effects on the behavior of poor people have often reemerged as concerns, including in the debate in the US in the 1980s and 1990s; Murray"s (1984)  "One side of American politics considers the modern welfare state-a private-enterprise economy, but one in which society"s winners are taxed to pay for a social safety net-morally superior to the capitalism red in tooth and claw we had before the New Deal. It"s only right, this side believes, for the affluent to help the less fortunate. The other side believes that people have a right to keep what they earn, and that taxing them to support others, no matter how needy, amounts to theft." However, while these debates continue, it is hard to imagine hearing today the kind of complacency about poverty evident in (for example) the above quote from Colquhoun from any prominent public person anywhere in the world. There is clearly a lot less poverty in the world as a whole (as a share of its population), and this fact alone may well have made it harder to justify continuing complacency. There is also far more knowledge about poverty in the world. Granted, opinions still vary on the causes of poverty and on appropriate actions for fighting it. Some observers still point to behaviors of poor people as causes of their poverty, while others point to constraints beyond their control that prevent them from escaping poverty. But it would seem unlikely that anything more than a small minority of people today would accept that poverty is a "necessary and indispensible ingredient of society."

Historical clues to the change in attitudes
So sometime in the last 200 years a world free of poverty became imaginable to at least the literate population at large, and it also came to be accepted that there were public actions that could help achieve that goal. This change in attitudes to poverty is not well understood, and poses many questions. When did the change happen and why? Were there reversals-multiple switches in dominant public attitudes over time? Were the changes synchronous with other changes in mainstream economic and political thinking? To what extent were attitudes to poverty 14 and policy bundled with other views on the role of markets and the state-views that are also known to have changed over time? In particular, did resurgence of public enthusiasm for marketbased solutions to economic problems over governmental ones in the late 20th century come with less concern about poverty and inequality?
The literature offers some clues on when the change in attitudes might have occurred. Fleischacker (2004, p.4) argues that the modern concept of "distributive justice"-by which he means that "property is distributed throughout the society so that everyone is supplied with a certain level of material means"-only emerged in its crudest form in the last two decades of the 18th century, around the time of the French Revolution. In attempting to identify an originator of the modern idea of distributive justice, Fleischacker gives credit to the French revolutionary "Gracchus" Babeuf and the German philosopher Johann Fichte, a follower of Immanuel Kant. 17 However, Fleischacker (2004, p.80) also notes that it was a long time before this new idea of distributive justice was to get beyond the "margins of respectability." Indeed, Fleischacker argues that it was not until John Rawls (1971) that the idea of distributive justice that first germinated in the late 18th century was given a complete and precise formulation.
In the history of thought on poverty in English at least, the literature has attached some significance to the period around the end of the 19th century and early 20th. The historian Robert Webb argues that this was the time in England that it came to be recognized that poverty "could and must be eliminated" (Webb, 1974, p.384 various scattered examples of the modern idea of distributive justice through the 19th century, he also points to the turn of the 20th century as marking a significant change. The last few decades of the 19th century saw the beginning of modern scientific research on social problems such as poverty-research that influenced policy debates and efforts for social policy reform. New professional associations for the social sciences emerged. 18 So did new research. Charles Booth and Seebohm Rowntree documented the living conditions of England"s poor (in London and York respectively) in the late 19th century. Booth and Rowntree 17 In his discussion of early pioneers in social thought, Jones (2004) points to the Marquis de Condorcet, who advocated free universal basic education and equal rights for women. 18 In the US, the American Social Science Association was founded in 1865 for the scientific study of social issues. The American Economics Association was founded in 1885. The Royal Economics Society was founded in 1890 as the British Economic Association (receiving its Royal Charter in 1902). did pioneering measurements, setting poverty lines and showing how people lived relative to those lines. Their work attracted much attention. The English public was shocked to learn that one million Londoners lived below Booth"s poverty line of 21 shillings per week for a family; this was a frugal line-by my calculations it was equivalent to 1.5 pounds of good wheat per person per day. 19 Booth and his writings are thought to have been instrumental in Britain"s introduction of a public pension in 1908. And debates began about where there is more poverty.
For example, 15 years after Booth"s books appeared, Alfred Marshall was argumentatively pointing out that there was even more poverty in Germany than Booth"s figures suggested was the case in England; this was in response to his perception that "one of the few things which every German knows for certain about England is that there are a million people in London living in extreme poverty on the verge of hunger" (Marshall, 1907, p.12).
The poverty studies by Booth and Rowntree were influential. Robert Hunter followed their lead in estimating the extent of poverty in the US in 1904. 20 Booth"s approach influenced the development of quantitative sociology in both Britain and the US. 21 Peter Townsend"s (1979) empirical study of poverty in England some 80 years later clearly owed much to Booth and Rowntree. So too did the Chicago School of sociology that began studying urban poverty in the US during the 1930s.
The new sequence from evidence and debate to policy action on poverty was to be replicated often in the 20th century. In 1990 many people were also shocked to learn that there were about one billion people in the world living on less than $1 per day, at purchasing power parity (World Bank, 1990). 22 Such data, and (probably more importantly) credible poverty data for specific countries, have been the empirical foundation of domestic and international efforts to fight poverty since the 1980s. 19 Marshall (1907) estimates that 21 shillings was equivalent to three-quarters of a bushel of good wheat. At 13.5% moisture by weight a bushel of wheat weights 60lb according to the Wikipedia entry on "bushel." I assume a household of 4.5 people, which is the lower bound of the range 4.5-5 given by Booth (1993, Chapter 4) for the average size of working men"s families at the time. This calculation was based on household surveys for just 22 developing countries (Ravallion, et al., 1991). The $1 a day line was a typical poverty line for low-income countries. Since then there has been a massive expansion in survey data collection and availability, and refinements to the methodology; the latest estimates can be found in Chen and Ravallion, 2010).
While the type of quantification initiated by Booth and Rowntree has been crucial, credibly reported qualitative observations have also been influential. I have already noted the impact of Harrington"s The Other America on policy in the US. Many people were influenced by Harrington"s efforts to "describe the faces behind the statistics" (p.17). In an introduction to a 1993 reprint of The Other America, Irving Howe (1993, p.xii) describes the "central premise" of Harrington"s work: "…that if only people knew the reality they would respond with indignation, that if people became aware of "the invisible poor" they would act to eliminate this national scandal." Clearly Booth and Rowntree shared the same premise.
In economics around the turn of the 20th century, Marshall (1890) was asking in the opening pages of his Principles of Economics, "May we not outgrow the belief that poverty is necessary?" (p.2). Marshall went on much later in Principles (pp.594-9) to sketch policies for fighting poverty that were not just intended as short term moralistic palliatives but were driven by a recognition that persistent poverty was itself a constraint on wealth generation-clearly a very different perspective to the mainstream view in the economics of 100 years earlier. 23 While Marshall was careful to avoid naïve utopianism (see, especially, the comments in Marshall, 1907), his writings reflect a far more positive perspective on social policy, as a means of expanding opportunities for all to share in the potential of a competitive market economy.
Did these new views also mark a broader change in public awareness and attitudes? Did that continue into the late 20th century? The enfranchisement of the poor that came with universal suffrage was largely a feature of the 20th century. This must surely have encouraged greater attention by governments to social issues, including poverty, though it is unclear if this surfaced in the form of broader public awareness or concern. By the mid 20th century the poor were probably a minority in Western Europe and America. Some observers have argued that the American public of the late 20th century was less interested in poverty and inequality than it used to be. Galbraith (1958Galbraith ( , 1998 wrote that "…few things are more evident in modern social history than the decline of interest in inequality as an economic issue." In the introduction to his 1975 text, The Economics of Inequality, Tony Atkinson noted how little attention economists have given to inequality. Similarly, Michael Katz (1987, p.7) described 23 Marshall gave emphasis to education policies that would make the children of the unskilled skilled. "The inequalities of wealth, and especially the very low earnings of the poorest classes..(are)…dwarfing activities as well as curtailing the satisfaction of wants" (Marshall, 1890, p. 599). what he saw as the "silence about poverty" in the US. This is, however, hard to reconcile with Figure 1; the rising attention to poverty seen in the Viewer after 1960 is hardly suggestive of a silence about poverty (though an even louder voice may have been needed).
Few of these anecdotal observations from the historical literature are borne out by Figure   1. While there is no doubt that the late 19th century was significant in the emergence of scholarly research on poverty, there is little convincing sign of heightened public awareness around that time. (There is at most a small positive blip around 1900 that might reflect, but it can hardly be deemed significant, as similar blips occurred at other times).
However, the changes in public awareness of poverty in both Poverty Enlightenments were undoubtedly grounded on less popular, more scholarly, writings over many decades. Other factors were no doubt also at work. One can speculate that fears of revolution may have led to efforts in popular writings to justify poverty and inequality in the late 18th and early 19th centuries-efforts that Landauer (1959) argued helped stimulate the development of socialism.
The spread of literacy may well have helped in generating a more empathic literature by the time of the Second Poverty Enlightenment.
It must also be realized that the time of the First Poverty Enlightenment lacked the body of theories and data that we take for granted today in trying to understand poverty and link that understanding to public action. Economics was still in its infancy. Systematic sample surveys of household livings standards were unknown. Nor was there much sign yet of the theories and movements that could claim to represent the interests of poor people. All that had changed by the Second Poverty Enlightenment, such that authors like Harrington and Galbraith could formulate accessible arguments grounded on a body of theory and data, the latter stemming from sample surveys and much analytic work in measuring living standards and setting poverty lines.
While tracking words over time cannot resolve all these historical issues, some further clues can be obtained by using the Viewer to track other n-grams related to poverty (including "inequality") and policies for fighting it.

"Poverty," "inequality" and related ideas in the literature
Inequality is often linked to poverty, in both theory and popular thinking. Figure 2(a) adds "inequality" to the series in Figure 1 from the English corpus, while Figure 2(b) gives "pauvreté" and "inégalité," drawing on the French-language books (about 10% of the words). References to "inequality" in English writings have been less frequent than "poverty," but there are similarities in the time profiles of the two words. There has also been a marked upward trend in references to inequality in the English literature since 1960. This is not consistent with Galbraith"s view of the declining interest in inequality (though possibly the enormous influence of Galbraith"s The Affluent Society changed matters). However, recall that Galbraith was writing specifically about the US; later I will focus on American publications only, to see if this offers more support for Galbraith"s view.
In French, the latter half of the 18th century was when references to "inégalité" peaked, once around 1760 and again around 1780, in the lead up to the French Revolution of 1789-99.
References to inégalité rose noticeably in the period 1830-50. This was the period of the "July Monarch" (the monarchy had returned to power after the defeat of Napoleon at Waterloo in 1815), which was seen to favor the wealthy merchant bourgeoisie; there was much criticism from the left at the time. After that, references to inégalité saw a trend decline until just before the First World War (WW2 hereafter), after which there was an increase, though uneven, and clearly less pronounced than in English.
Both "Poverty Enlightenments" are evident in French, and coincide pretty closely with those in English. The two words "pauvreté" and "inégalité" were only weakly co-moving in the 18th century. They were used with similar, and declining, frequencies until the end of the First World War (WW1) but diverged after that. Eventually the pattern in their usage became similar to that in English, though with a marked decline in references to "inégalité" starting around 1980. Comparing panels (a) and (b) of Figure 2, we also see that the incidence of references to poverty is considerable higher-about three times higher in 1800-in the English language than in French. The opposite is true for inequality. But this difference too was much reduced by 2000.
In both languages, the moving average incidence of references to poverty peaked at or near the end of the period. 24 This was almost certainly the time when the incidence of absolute poverty in the world as a whole had reached its low point, at least when judged by what "poverty" means in developing countries today. 25 This is consistent with the view that, as poverty falls with overall economic development, the capacity of an economy to address the problem of poverty increases (as argued in Ravallion, 2010), which probably attracts attention to the problem in the literature. Changes in the frequency of public discussion of poverty might be 24 As already noted the single year with highest incidence was 1787, but this was a spike, and not the moving average peak. 25 As such measurements require survey data, this claim can only be tested for the last couple of decades; see Chen and Ravallion (2010). Bourguignon and Morrisson (2002) provide simulations suggesting that the poverty rate declined from 94% in 1820 to 51% in 1992. the natural outcome of success against poverty, even if there was little change in the distribution of social preferences related to distribution. 26 The rise in incidence of references to inequality in the English language books since the 1950s coincides roughly with the time when, according to the estimates by Francois Bourguignon and Christian Morrison (2002), the overall level of inequality among persons in the world started to stabilize, after a long-run increase going back to the early 19th century. So here too, public attention in the general literature is not in tune with the measure of the extent of the problem. However, there are a number of measurement issues that one would need to consider in a complete assessment of this issue. These include whether the relevant measure of "inequality" is that between all people in the world, or the average level of inequality in the main Englishspeaking countries; by that definition, inequality has been generally rising in Britain and the US over much of the period in which the incidence of references to inequality has also been rising. 27 Another observation supporting my interpretation of the "1970 bubble" in references to poverty is that there is no sign of this bubble in French (Figure 2(b)). Nor is it found if one uses the corpus for "British English" (books published in the UK) though it is strongly evident using "American English" (books published in the US) for which the bubble also marks the overall peak in references to poverty, and it appears slightly earlier, around 1970 ( Figure 3). The "1970 peak" in references to poverty coincided with the rediscovery of many other social problems in the late 1960s, notably (but not only) in America. Focusing on just the 20th century and removing all smoothing, Figure 3 also plots "crime" and "pollution"-which also peak around 1970 in the American literature, but not the British. Like poverty, references to "crime" also peaked in the 1990s. In their short-term fluctuations over time, these two words are not strongly co-moving-rather it is these major peaks that they have in common.
Figure 3 also gives the series for "inequality" in both American and British literatures. In both cases there is rising incidence of references, and the rise came earlier in America, although the peak (near the end of the period) is at a higher incidence in Britain. 26 For an attempt to assess how the capacity for addressing poverty through redistribution changes with economic development see Ravallion (2010). 27 See, for example, Lindert (2000). There have been a number of other expressions for "poverty," including "indigence," "pauperism," "destitution," and "penury." 28 These are often interpreted as extreme forms of poverty. These have generally seen declining usage over the period, as can be seen in Figure 4.
There was increasing use of the word "destitution" in the first 40 years of the 19th century, which made up for the declining use of "poverty" (as can be seen more clearly here).
However, if one adds up these terms with "poverty" then the overall downward trend until 1960 in Figure 1 is even more pronounced, as is evident when the various words are plotted together 28 "Misery" is another term, but it is too ambiguous; the graph is here.
with "poverty," as can be seen here. So the conclusion that there was a trend decline in attention to the topic until around 1960 is robust to allowing for these other words for poverty.

Figure 4: Other words for poverty
The way "poverty" is referred to can reveal something about how the writer thinks about poverty. Depending on the context, references to "the poor" can carry an impression of social distance, and little empathy, while a term such as "poor people" is more neutral. References to "the poor" have been declining over time ( Figure 5), consistent with the view that writers in the 19th century were not as empathetic with poor people. Granted, a 1970 surge in use of this term is evident, coinciding with that for "poverty" (Figure 5), but there is no sign of the sustained increase we have seen since 1980 in the references to "poverty" and references to "the poor" continued their long-run decline. With changing attitudes to poverty, the phrase"s somewhat stiff and pejorative tone has probably contributed to this long-term decline.
A long-standing distinction is between "absolute" and "relative" poverty. There is some ambiguity in usage of these terms in both English and French. Absolute poverty sometimes means "extreme poverty" while sometimes it simply means that the poverty line is fixed in real terms. Relative poverty sometimes refers to one person"s poverty relative to another, and sometimes to a situation in which the poverty line varies in real terms, such as being set as a constant proportion of average income. 29 There is also a variation across space and time in the 29 The weight attached to relative poverty has often been politically charged, which has clouded the measurement issue. Indeed, it can be argued that neither the extreme forms of absolute poverty nor real value of the poverty line, with higher real poverty lines prevailing in richer countries; also sustained economic growth tends to put upward pressure on the poverty line (even if it is fixed for even quite long periods), as the standards used to define "poverty" evolve. 30 Booth"s poverty line (which was frugal indeed, as already noted) would never be accepted as a meaningful characterization of what poverty means in London today. Absolute poverty was used more frequently than relative poverty in English books until the 1930s (Figure 6). The graph is similar between American English (here) and British English (here). Relative poverty was the more common term for the next 40 years or so. A dramatic increase in references to "absolute poverty" in the 1970s peaked in the early 1980s, though with signs of convergence with relative poverty in the 1990s.
French texts share a similarity in the reversal in usage of these two terms, but "pauvreté relative" dominated from much earlier. The same dramatic increase in attention to "pauvreté absolue" occurred in the 1970s, continuing through to the "90s. Relative poverty was the dominant concept in French for many more years than in English. By 2000, the share of all words represented by absolute poverty was about the same between the two languages. relative poverty has a credible welfare-theoretic basis, though an intermediate concept of "weakly relative poverty" does satisfy a number of plausible assumptions . 30 Evidence in these claims can be found in Ravallion (2011). Class consciousness peaked around the 1970s. References to the "lower class" dominated until the mid-1840s, after which references to the "middle class" dominated, and diverged markedly from "lower" and "upper" class ( Figure 7). References to all three, and "social class," peaked around 1970, and declined since. References to "working class" followed a similar pattern but peaked later. So the Second Poverty Enlightenment was not one of heightened class consciousness, but heightened poverty consciousness.  The extent to which poverty is seen as being caused by the behavior and circumstances of poor people, versus broader social factors outside their control, has long been a theme in the literature. As can be seen in Figure 9, "feeblemindedness" rose to some prominence as an explanation of poverty in the early part of the 20th century, notably with Henry Goddard (1914), 26 who argued that mental deficiencies were a cause of poverty among immigrant families in the US. The term declined from favor from about 1950, although an effort to attribute poverty to lack of intelligence resurfaced in the 1990s with Richard Herrnstein"s and Murray"s (1994) The Bell Curve, who argued that intelligence trumped family background or education as predictors of poverty in the US. This book was a best seller and the subject of a great deal of controversy, particularly (but not only) for its claims about the association between race and intelligence. 31 The ideas of a "culture of poverty" and an "underclass" that emerged in America in the 1960s also identified various behaviors by poor people as the causes of their poverty. These ideas were controversial. Herbert Gans (1995) and Alice O"Connor (2002) criticized them for ignoring more deep rooted "structural" inequalities. 32 In some versions of the "underclass" idea, such as in William Julius Wilson"s (1987) The Truly Disadvantaged, a "culture of poverty" was seen to be rooted in structural inequalities and so part of their explanation, 33 but even so there has been much debate about the role of cultural factors. The literature of the late 20th century came to question whether there was any legitimate space for policy intervention in changing culture. 34 The Wikipedia entry on The Bell Curve discusses the various criticisms at the time.

32
The idea that poor people are to blame for their poverty appears to be more common in Anglo-Saxon countries; see Van Oorschot and Halman (2000). 33 Wilson emphasized the importance of macroeconomic factors, including structural changes in the economy, urban structural changes and aggregate unemployment rates. 34 This debate re-opened again in Steinberg"s (2011) comments on Small et al. (2010), also echoing his past debates with Wilson. Use of the term "underclass" rose through the 1970s and 1980s, but peaked in the mid-1990s. The idea of a "culture of poverty" started to fall out of favor earlier, from about 1970 ( Figure 9). 35 References to "welfare dependency" rose in the 1980s and (particularly) 1990s, coinciding with the (highly politicized) debate on welfare reform in the US in the mid-1990s.
References to the "working poor" were more common, and also rose from the 1970s (Figure 9). This was happening while attention to the "working class" as a concept was abating.
The English Poor Laws back to the early 17th century had been premised on the idea that there were both "deserving" and "undeserving poor," usually defined according to the ability to work, and that public relief should focus on those unable to work. (Indeed, according to the Peel Web, under Elizabethan Poor Law, the able-bodied paupers were "severely beaten until they realized the error of their ways".) Much later in the 20th century, there were concerns about "welfare dependency"-that the provision of welfare itself left little incentive for poor people to take actions that would lift them out of poverty. 36 Social surveys have suggested that whether welfare recipients are considered "deserving" influences public support for welfare payments. 37 While antipoverty policies have tried often to target the "deserving poor," the meaning of this distinction is unclear in the models of poverty that show how the behaviors that make one seem "undeserving" have deeper structural roots; then the label risks adding insult to injury. 38 Figure 10 plots both terms. (There were no references to either before 1790.) References to the "deserving poor" peaked around 1900, but there has been a recent resurgence of it use since 1960, which was sustained alongside a marked increase in references to the "undeserving poor." The steep increase in references to the undeserving poor in the 1990s came alongside increasing emphasis on individual responsibility in welfare-policy discussions. As already noted, incentive effects have long been an issue for critics of direct interventions against poverty, ranging from Thomas Malthus to Charles Murray. 37 See, for example,  and Applebaum (2001). 38 This point was made well by James Boswell (1791, p.217) in his Life of Samuel Johnson, "A particular kind of manufacture fails; those who have been used to work at it, can, for some time, work at nothing else. You meet a man begging; you charge him with idleness; he says, "I am willing to labour. Will you give me work?' 'I cannot.' 'Why, then you have no right to charge me with idleness.'" Historically, beggars, vagabonds and vagrants were more likely to be sent to jail than the poor house. The 19th century had a far higher incidence of references to them (Figure 11). 39 The decline in use in the 20th century was interrupted by the Great Depression. Since 1980 there has been a huge increase in references to the "homeless."

Figure 11: Hobos and homeless
The distinction between urban and rural poverty has been especially important in developing economies, where rural poverty tends to be more prevalent, but significant biases have existed in development policy against the rural sector; Michael Lipton (1976) was influential in efforts to redress these biases. It also came to be realized that the process of 39 The plurals of these words track their singular forms closely, so I dropped them from the graph. economic growth through sectoral transformation (and urbanization with it) entails a pronounced "urbanization of poverty" (Ravallion et al., 2007). 40 We find in Figure 12 that "rural poverty" has received more attention than "urban poverty." The 1960s saw a sharp increase in references to both, but the period 1970-90 saw a small decline in interest in urban poverty, while references to rural poverty grew markedly; this pattern was reversed around 1990. The signs of relatively more attention to urban poverty are consistent with the evidence on the urbanization of poverty in developing countries over the last 20 years or so of the 20th century (Ravallion, et al., 2007).

Figure 12: Urban and rural poverty
While references to "poverty" and "inequality" go back centuries, their measurement has been confined mainly to the latter half of the 20th century when household survey data became common. Figure 13 plots the most commonly used terms in measurement. 41 The term "poverty line" goes back to the late 19th century, and appears to have been first used by a reviewer of Booth"s study of poverty in London. 42 However, the bulk of the growth in usage was much later, during 1960-90. (Here is a "blow-up of the period 1960-2000.) References to the "poverty rate" grew rapidly from 1980, and continue to the end of the period.

40
See, for example, the discussion in Lipton and Ravallion (1995). On the evidence see Ravallion et al. (2007). 41 My initial selection included "headcount index" and "headcount ratio" but they are used with low frequencies so I dropped them. 42 The earliest reference I can find to a "poverty line" is in a description of Booth"s poverty line for London in the Speaker (1891). This predates the earliest citations to this concept found in Gillie (1996). References peaked in 1980, but there was a similar peak in the late 1990s.
The extent to which poor people share in the gains from aggregate economic growth has long been debated. It was thought likely that growth would increase inequality in the early stages, but possibly reduce it later (following what has come to be known as the Kuznets Hypothesis, following Simon Kuznets, 1955). Subsequent evidence has not borne out this view.
Inequality has decreased roughly as often as it has increased in growing developing economies, and the incidence of absolute poverty has tended to fall. 45 Figure 14 plots "economic growth" and a set of other n-grams. (A blow-up for just the 20 th century can be found here.) We see in Figure 14 that the phrase "economic growth" appeared on the scene around 1950, though it was clearly implicit in much earlier writings. This did not, however, displace attention from distributional concerns. "Inequality" is an older term as we have seen. It is notable how much interest in inequality has kept up with interest in economic 43 A small minority of the references are to "Gini" as another word for Guinea, where gold was mined. However, it appears that the vast bulk of the growth in usage after 1950 was for the Gini index. Note also that the Gini index is not only used in economics to measure inequality, but is also used in the physical and biological sciences and engineering. 44 Atkinson"s (1970) paper was clearly a landmark in the outset of this resurgence of interest. 45 For a recent overview of the literature and evidence see Ferreira and Ravallion (2009). growth. There is no sign here that concern about inequality has abated with the rising prominence of economic growth as a topic in the literature.

Figure 14: Other economic and social issues
Concerns about inflation and unemployment-central concepts in macroeconomics, also known to have bearing on poverty-have fluctuated. Unemployment clearly existed well before the word appeared around 1900. Unsurprisingly, references to unemployment rose sharply in the wake of the Great Depression and peaked around 1940, abating equally rapidly until the mid 1950s. The next peak was in the 1980s; US unemployment rates reached 10% in the early 1980s, along with high inflation rates. Only the latter peak coincided with that for inflation; there was little concern about inflation around the time of the Great Depression, but the two came hand-inhand in the form of "stagflation" in the 1970s, which clearly stimulated the rising attention to both terms seen in Figure 14. Note also that the far higher unemployment rates (over 20% in the US) in the Great Depression generated a lower incidence of references to unemployment than did the recession of the early 1980s. The recent literature seems more responsive to these events.
"Economic development" emerged as an idea around 1900 (Figure 14). It gained usage slowly until 1950, after which there was a sharp increase; indeed, its share of all words increased by a factor of five in the 20 years after 1950, though stabilizing after 1970. The term "Third World" appeared around 1960, ballooned in the next 20 years, but faded rapidly in the 1990s.
"Famine" has long been linked to poverty (a linkage made more rigorous by Sen, 1981.) However, while references to "famine" were common in the early 19 th century (the incidence 32 peaked around 1770, as can be seen here), and declined similarly to "poverty" in the following 150 years, interest in famine did not see the same resurgence in the latter part of the 20th century as did poverty. The incidence of the word "hunger" has stayed fairly constant over time, and shows no sign of the marked increase in attention to poverty after 1960.
References to "migration" rose in prominence through most of the period, though with a brief waning of attention in the 1980s. Behaviors still play a role, such as when poor parents are compelled to take their children out of school to earn extra current income. But this is put in a broader economic context, noting the credit market failure that prevents poor families from borrowing on the basis of their childrens" future earnings from better schooling.

Schools of thought
Classical economic theory has often been criticized (at the time and since) for taking the distribution of wealth as largely outside its domain. 46 As discussed above, this came with a perceived acceptance of poverty and inequality (though class-based inequalities received attention by critics of capitalism in the 19th century). In the 20th century, while poverty has been a subject of study in all social sciences, economics has clearly been the main discipline linking poverty to policy. Indeed, it has been said that "Economics is, in essence, the study of poverty" (Hartwell, 1972, p.3). So before we start looking at policies, it is of interest to look at the evolution of references in the literature to various ideas from political economy and economics. The term "economics" first appears in the corpus with more than negligible frequency around 1880, though the term "political economy" had been used well before that ( Figure 15). 47 There was a local peak of interest in political economy around 1890. A more pronounced peak in references to economics, was in the late 1930s, in the wake of the Great Depression, which clearly stimulated economic thinking, though attention was then deflected by WW2. Mentions of economics became more frequent from 1945 onwards. It is striking that it took about 25 years for economics to return to the attention it had just prior to WW2. "Political economy" saw a 46 See for example Maurice Dobb (1975). 47 "Expediency" was also used to mean economics in the late 18th century and early 19th. But the frequency of usage was low, and I dropped it from the graph. resurgence of interest in the 1990s though with more specific meaning, emphasizing the political.
However, there is also a sign that attention to both topics peaked in the mid 1990s.
The rise of socialism in the 19th century may well have attracted writers and readers concerned about poverty, which might explain the declining references to "poverty" from the mid-19th century onwards. Marxism attached little explicit weight to concerns about poverty and inequality, 48 but instead focused on the alienation of workers from their output that was seen to be a defining characteristic of capitalism; distribution was not seen as separable from how production was organized. Instead of referring to "poor people," one talked about the "working class." Does this explain the declining in attention to poverty in the latter half of the 19th century?
We know from the literature that the idea of "socialism" goes back to the late 18th century (Landauer, 1959). References to the word "socialism" in the corpus go back to that time, but with very low frequency. The Viewer graph does not suggest that socialism go much attention in the first half of the 19th century ( Figure 16). The literature has identified a crucial point in the history of socialism as being when, sometime around the middle of the 19 th century, its philosophical side merged with the labor movement (Landauer, 1959). However, there is no sign of a large increase in public awareness of socialism at this time. Indeed, the incidence of references to socialism remained quite low until about 1880, after which we see a large increase in references to the word and (with a lag) related ideas ( Figure 16). 49 Most of the series in Figure   16 peaked in the 1970s ("Marxism" later). These trajectories bear little obvious relationship to the time profile of references to poverty and inequality (Figure 2(a)). The initial upsurge in attention to poverty in the literature in the 1960s and "70s did coincide with rising attention to Leftist political ideas, but the two went separate ways after 1980. It is clearly not the case that the trend decline in references to "poverty" in the 19th century reflected a displacement of attention toward socialism or the working class. 48 The word "poverty" appears only 20 times in Marx (1867) and rather tangentially to the main argument. 49 I dropped the word "Communist" as its usage far exceeds the other terms (making it hard to read others), but it follows a similar pattern, rising until the 1960s, falling after that (as can be seen here).  The "Chicago School" and "Cambridge School" were on a par in the 1950s, but then the former took off on a steeper trajectory; mentions of the Cambridge School decline from the mid-36 1980s, as did the Chicago school from about 1990 ( Figure 17). 50 References to "Keynesian economics" peaked in the mid-1980s. The term "neoclassical economics" appeared on the scene around the end of WW2 and saw a steep rise in attention from around 1980, though with a marked reversal around 1995. "Institutional economics" also saw rising references from 1980s onwards.
A striking similarity between the two Poverty Enlightenments is in the prominence given to rights in thinking about how society should be organized, as can be seen from Figure 18 (compared to Figure 1). Discussions about rights in the latter part of the 18th century were often about how best to reconcile individual liberties with membership of a society, and the role and powers given to government to constrain liberties and confer rights. Poverty was not central in such discussions, but it was present, including in the writings of Kant (1791)-one of the great advocates of the primacy of individual rights (consistent with the same rights for all), who also defended governmental assistance for poor people as an element of the social contract. Kant"s equivalent in the Second Poverty Enlightenment was Rawls (1971). These concepts of a "just society" emphasized the importance of assuring that poverty itself does not constrain the fulfillment of individual potential-an idea that has been developed in the "post-Rawlsian" writings of Amartya Sen (1985) and others. By contrast, utilitarian and socialist philosophies played down, or even dismissed, concerns about individual rights (Fleischacker, 2004). There is some ambiguity here, as there is also a Chicago School of sociology, which predates the Chicago School of economics. (Wilson, 1987, was a product of the Chicago School of sociology.) Many of the early references to Chicago School in Figure 17 clearly relate to sociology not economics.
Welfare economics has often been employed in addressing issues of poverty and inequality (though not always explicitly). Here Sen"s (1979) distinction between "welfarist" and "non-welfarist" is relevant, since important conceptual ideas related to poverty (such as "basic needs") are more easily understood in non-welfarist terms. Figure 19 plots selected terms. (The word "welfarist" has not, however, come into more than negligible usage in the general literature.) Figure 19: Welfare economics and distributive justice The term "social justice" is largely confined to the 20th century, during which there was a large increase in usage, albeit with (puzzling) attenuations in interest in the aftermath of both WW1 and WW2. The term "distributive justice" is far older, though arguably had a somewhat different meaning in earlier times, with a modern usage closer to the idea of "social justice" (Fleischacker, 2004), and there are signs of co-movement after WW2. There was a slowdown in the rising attention to social justice, distributive justice and "Rawls" in the 1980s, but resurgence in the 1990s.
"Utilitarianism" has a long history, of course, going back to Jeremy Bentham in the late 18th century and developed further in the 19th century by John Stuart Mill, who had enormous and lasting influence on economics and policy. References to "utilitarianism" peaked 200 years after the publication of Bentham"s Principles of Morals and Legislation. The term "social welfare" grew rapidly in usage in the 20th century, peaking around 1980, but staying in reasonably steady use since then. Some of the usages of "basic needs" plotted in Figure 19 (such as during WW2) have somewhat different meanings to the use that started to grow rapidly in the 1970s, associated with concerns about poverty in developing countries. This usage also peaked around 1980 and the term has fallen somewhat from favor, though with "human development" making up some of the difference; the acceleration in use of the latter term since the Human Development Reports appeared in 1990 is evident (UNDP, 1990).

The role of the state and public policies
While mainstream economic thinking of the 19th century provided strictly limited support for government intervention, that changed radically in the 20th century. The large increase in public spending (as a share of national income) in that century is generally ascribed to the turnaround in thinking about the role of government in the post-WW1 period. The post-WW2 period up to about 1980 has been dubbed the "golden age of government intervention" by Vito Tanzi and Ludger Schuknecht (2000). A broader role for government and wider range of interventions emerged in this period, including greater acceptance of the governmental social protection. The last 20 years of the 20th century then saw greater caution about the efficacy of government interventions in general, and stronger political movements calling for a more restricted role for government, echoing the mainstream views of the 19th century.
We saw in Figure 1 that references to poverty fell in the early 1970s. Was this associated with the push-back on government intervention in favor of free-market solutions to economic problems? This seems unlikely for two reasons. Firstly, this happened before the economic and political counterattack on government intervention was taking hold, which is generally thought to have been closer to 1980. Secondly, there was no sign of this in the series for French-language books (Figure 2(b)) or in the series for British English (Figure 3(b)). It appears more plausible that the decline in references to poverty in the early 1970s was a separate dynamic, built on the (rude) awakening of America"s middle class and rich to the existence of the "Other America" and subsequent return to its normal trajectory. The aforementioned fact that the success of Harrington"s book was a surprise also supports this interpretation.
References to some indicators of "intentions" for action against poverty and inequality are plotted in Figure 20. The surge in mentions of "antipoverty" in the 1970s appears to have originated mainly in the US, and the word "antipoverty" was often followed by "programs;" the term is far less common in the British English corpus, as can be seen here. "Poverty alleviation" grew rapidly in usage in the 1980s, peaking in 2000. References to "redistribution" peaked around 1980, and fell out of favor in the 1980s and 1990s. "Redistribution of wealth" was often mentioned in the Great Depression, but declined in use during WW2 and after until about 1960. Turning to more specific policies, these can be considered at three levels: broad economy-wide policies, generic types of social policies, and specific social policies. Figure 21 provides selected terms under these three headings. (I modified my classification of topics somewhat to try to avoid grouping very low-count items with high-count ones.) Within the 20th century-here is the blow-up of Figure 21(a))-we first see a rise in attention to issues such as "nationalization" and "economic planning," 51 but this peaked in the 1970s, after which we see a marked increase in the incidence of references to "privatization" (dramatically so in the 1990s), "market economy," and a resurgence of references to "free trade" (Figure 21(a)). The incidence of the term "free trade" tended to grow in the 19th century, peaking around 1890; the frequency fluctuated without a clear trend in the 20th century up till 1980, when it saw a sharp rise, peaking in the mid-1990s. "Industrial policy" peaked in the mid-1980s, and followed an almost identical course to references to "protectionism"; these two ideas have clearly been strongly linked. References to "globalization" grew rapidly in the 1990s. 51 "Central planning" followed a similar course to "economic planning" up to 1970, but peaked later, in the 1990s. Clearly much of this was criticisms of central planning. The term "social security" emerged in the 1930s, grew rapidly, and stayed prominent for the rest of the period. References to "social policy" and the "welfare state" continued to grow rapidly through the late 20th century. Figure 21(a) does not suggest that the push-back against government intervention that emerged in the late 1970s came with reduced public concern about poverty and inequality, at least as reflected in references to these terms in the literature. References to "socialism" and "nationalization" started to decline from the 1970s, while references to "privatization" and "free trade" increased-along with references to poverty and inequality.
The second half of the 20th century saw an explosion of interest in a wide range of domestic social policies. Figure 21(b) turns to specific types of policies, and here is a blow-up of the 20th century only. Appeals to universal education go back to the early 19th century but peaked in the 20th. Most of the direct antipoverty policies only came to public attention in the 20th century. We see that "progressive taxation" was the most frequently used of these terms up until about 1960, but then attention switched to the public spending side of the budget.
The idea of a "means test"-whereby transfers are targeted to those with low measured incomes-saw an uneven rise in attention until the 1970s, but lost ground after that. There was a steady rise in references to "basic income." There have been various proposals for a "basicincome" (also called a "guaranteed income" or "citizenship income"), which provides a uniform (un-targeted) transfer to everyone, whether poor or not. 52 There was a sharp increase in use of "social spending," "social protection," "public pension," and "workfare" from about 1980. Of course, references to these terms were both positive and negative; this was also a period of welfare-policy debate, such as in the US in the 1990s. The 20th century clearly saw a change in attitudes, and a broader set of policies were seen to be relevant to poverty. There were references to the Poor Laws and the "workhouse" going back to 1800. The first minimum wage law was introduced by New Zealand in 1894. References to "minimum wage" 52 Raventós (2007) has been a prominent advocate of these policies. fluctuated considerably in the 20th century, falling during the Great Depression and WW2, but rising steadily from the early 1950 to peak around 1980. References to Franklin Delano Roosevelt"s "New Deal" rose sharply from the early 1930s to peak in the mid-1940s.
The US program, "Aid to Families with Dependent Children" (AFDC) actually started in the 1930s, but only received significant public attention from 1960. AFDC expanded from then until the mid 1990s, though real benefit levels per recipient had been declining since the mid-1970s (DHHS, 2008. AFDC was at the center of the debate on welfare reform in the US in the 1990s, and was disbanded in 1996 (replaced by a more restrictive program).
Another notable change in the late 20th century was the rise in public attention to civil society groups, including non-governmental organizations (NGOs). This can be seen in Figure   22. Charities have a long history back to mediaeval times (as documented by Geremek, 1994).
There was a decline in incidence of references to "charities" and "benefactions" from around 1820. From the early 20th century, this came with rising references to the terms "nonprofit" (from 1950s) and "civil society" (from around 1980). To some extent these are different labels for similar things, but it is also evident that the aggregate incidence of references to all of these words has increased markedly in the late 20th century.

Conclusions
There have been significant changes in attitudes to poverty over the last three centuries, away from complacent acceptance, and even contempt for poor people, to the view that society, the economy and government should be judged in part at least by their success in reducing poverty. There are a number of possible explanations for this change. Greater overall affluence in the world has probably made it harder to excuse poverty. Expanding democracy has given new political voice to poor people. And new knowledge about poverty has created the potential for more well-informed action.
Word counts from Google Books suggest that there was a pronounced "Poverty Enlightenment" in the latter half of the 18th century, with a seven-fold increase in the incidence of references to poverty between 1740 and 1790. The all-time annual record level of references to poverty (as a proportion of all words) was in 1787.
But attention faded after that. If the late 18th century gave birth to the modern idea of distributive justice, based on the notion that a minimum standard of living should be attainable by all members of society, then it seems that the idea died a slow death in the public consciousness for the next 170 years.
In both English and French, a second "Poverty Enlightenment" only emerged in the latter part of the 20th century. Within 10 years the incidence of the word "poverty" in Englishlanguage Google Books had exceeded its level even in 1800. The latter half of the 20th century saw a doubling in the incidence of references to poverty. Averaging over a few years either side, the peak in the average incidence of references to poverty was around 2000. This Second Poverty Enlightenment came with a similar (proportionate) increase in references to economics, which provided a deeper set of models for understanding poverty. It also came with rising awareness of poverty in developing countries.
In the English language, the word "inequality" has been used far less than "poverty." This was not true in the French language books. Until the late 20th century, "inégalité" was as prominent as "pauvreté" (more so in the late 18th century). And the French texts referred less to poverty and more to inequality than the English texts, though there are strong signs of convergence between the two languages toward the end of the 20th century.

44
The last 300 years have seen large swings in attitudes toward markets and the state, including on the potential for effective government intervention. The post-WW2 period-the "golden age of government intervention"-saw rising attention to a wide range of policies. The counterattack-based in some measure on economic analyses of the limitations of governmental solutions to economic problems, but also driven by an effective political mobilization-started to gain ground from the late 1970s.
However, while the cycles of debate and reform continue, the written record does not suggest that the end of the golden age of government intervention came with a widespread diminution of interest in poverty. Indeed, the incidence of references to poverty (and inequality) followed a clear upward trend in the latter half of the 20th century and, since 1980, we have seen a steep increase in references to social policies, social protection and civil society organizations-no doubt (in part at least) as a reaction to rising concerns about poverty and inequality. Similarly, the rising attention to economic growth-which was rarely mentioned explicitly before 1950-came hand in hand with rising attention to inequality.
The finding that the share of language devoted to both poverty and anti-poverty policies peaked near the end of the late 20th century can be contrasted with the emphasis historians have given to earlier times. The literature has pointed to the role played by new thinking on justice in the late 18th century and the emergence of scientific research on poverty in the late 19th century.
There were clearly formative times in the (long) evolution toward more progressive ideas on poverty. Today"s economic and political debates on poverty and policy echo those of the past. However, the time period since 1800 that has the strongest claim to have been a second "Poverty Enlightenment" is the late 20th century.
Translating this peak of awareness at the outset of the 21st century into effective action against poverty is, of course, another matter. The Poverty Enlightenment of the late 20th century was one of much debate and a mixed record of successes and failures in the fight against poverty.
But it is at least encouraging that the recent resurgence of the resistance to government intervention that prevailed through most of the 19th century does not appear to have come with a return to that century"s complacent acceptance of the inevitability of poverty.