Posted on

Conflating perception with judgment: the example of the Loudun Trial (1633-1634)

On a frigid December morning in 1633, Urbain Grandier, a catholic priest living in the French city of Loudun, is arrested at his home. The magistrate who has brought him the arrest warrant, the baron de Laubardemont, is the King’s special envoy to investigate what will be later considered as one of the most notorious case of witchcraft in the history of France.

The plot began a year earlier, in the Ursuline convent of the town, where a few of the seventeen nuns and the prioress Jeanne des Anges, started to feel unwell. The illness was quickly labelled as one affecting the mind, because many of the nuns were behaving oddly: they were strolling aimlessly around at night, crying frequently for no apparent reason, and expressing unusual desires. The prioress herself, Jeanne des Anges, said she was suffering from delusions and nightmares, and that her novices witnessed disturbing apparitions within the priory. The name of Urbain Grandier would be finally uttered, and the ecclesiastical jurisdiction alerted.

The behaviour of the nuns and the ‘unnatural’ scenes the investigators would witness the following days inevitably led to the opening of a case against Urbain Grandier for allegation of witchcraft. After a long and procedural investigation (the file amassing a weight of not less than 4,000 pages), Grandier would be condemned to death by burning. The unfortunate was executed the 18th of August 1634.

In these times, France was slowly recovering from recent wounds: the Reformation profoundly marked the country, and the King, a catholic, sought to regain control over the cities still influenced by the Huguenots. First and foremost, the wound was one of trust, and the reconciliation with the populace would only be far ahead. The city of Loudun was the stage of tensions and divisions, to the point of being threatened with destructive retaliation from the King himself. For the historian and philosopher Michel de Certeau, the Loudun possession unfolded at a time of psychological and social mutations, where science and its method were progressively replacing bigotry and superstitious knowledge (Certeau, 2005). The metamorphosis is illustrated by the very presence in the investigation panel of the Grandier case, of physicians and apothecaries along with representatives of the church and the crown. There was now a need for objectivity.

In this context, how could a trial based on fictitious allegations and superstitious precepts have led to such a tragic conclusion? The first reason is political, Grandier had a powerful nemesis in the person of the Cardinal de Richelieu, to whom the Baron de Laubardemont was plainly devoted. There is a second, more intriguing reason: a significant part of the jury – physicians included – was convinced of the magical extent of the case. By all means, jurors certainly relied on second-hand accounts of some shady witnesses, but the experts in the panel, the experts themselves, saw the devil at work. Or did they, really? Rogier, Cosnier, Carré and Duclos, the medical doctors in charge of the case, wrote, after attending an interrogation of the nuns in April 1634: “Nous avons jugé qu’il y a quelque chose qui dépasse la nature” (“We have judged that there is something beyond nature”). However, there is something deeply equivocal in the report of the physicians, something that could expand to and even explain the whole case: it is the conflation of perception with judgment. In the words of one of the judges, Mgr. de La Rocheposay “I’ didn’t come here to see if the possession is genuine. I already knew it is the case” (Certeau, 2005).

The Loudun trial remains an interesting example of a witness’s ability to bring his eyes to the point of lying. This apparent distortion between the percept and its cognitive processing might be considered as the basis of modern investigation in psychology and neuroscience: questioning the difference between believing and seeing, between what passes the retina and what sense one makes of it. Consequently, to understand how the mind makes sense of the world, we need to explore how it deals with uncertainty, what it cannot discern fully, what it needs to fill in the perceptive gaps with. More than two centuries after the Loudun possession, William James wrote, in his notorious Principles of psychology: “The brain is an instrument of possibilities, but of no certainties. But the consciousness, with its own ends present to it, and knowing also well which possibilities lead thereto and which away, will, if endowed with causal efficacy, reinforce the favourable possibilities and repress the unfavourable or indifferent ones” (p. 141, James, 1890). In James’ words, our actions are the product of our ability to distil possibilities into certainties, a singularly appealing perspective in the era of the Bayesian brain.

References

de Certeau, M. (2005). La possession de Loudun. (Gallimard, Ed.), La possession de Loudun (Folio Hist). Paris. https://doi.org/10.14375/np.9782070319138

James, W. (1890). The principles of psychology (Vols. 1 & 2). New York Holt.

Posted on

Goodness-of-fit test for the von Mises distribution: the bootstrapped Watson test in R

The von Mises distribution, a close approximation of the wrapped Gaussian, is a particularly popular option to describe circular, unimodal and symmetric data. The parameters of the distribution, the mean and the kappa, can be estimated using classical fitting procedure like Maximum Likelihood Estimation (MLE) or Bayesian estimations.

For model selection, one should be able to test whether the choice of a von Mises to describe a given dataset is sound. The use of a standard goodness-of-fit test like Kolmogorov-Smirnoff (K-S) or the Cramér-von Mises criterion (C-vM) is not recommended, because of the need for setting an arbitrary reference point on the circle in order to define the CDF. In other words, because the space is circular, one should expect invariance of the test results to space rotations, something not considered in the K-S and C-vM test. Two alternatives are usually used when dealing with circular data: the Kuiper and the Watson tests. Here, I propose a basic bootstrap implementation of the Watson test in R.

The U2 statistic and its asymptotic null distribution

Bootstrap script in R

Existing implementations of the Watson tests in R are part of the Circular and CircStats packages. They are using a critical points table for the asymptotic U2 distribution provided by Lockhart & Stephens (1985), which links different (Kappa,Alpha) ranges to U2 diagnostic values. However, this table is only providing significance ranges, and has an lower bound limit of 0.005 for the p-value. Here, we will try a very basic implementation of a bootstrap version of this test, where instead of the original table, we will be computing a bootstrapped U2 distribution to compare to our data. This way, p-values can be approximated at an arbitrary precision level.

First, we need to define our R function for the U2 statistic:

require("circular")

# Watson U^2 statistic
get.u2<-function(x){
    n <- length(x)
    res <- mle.vonmises(x, bias=FALSE)
    mu.hat <- res$mu
    kappa.hat <- res$kappa
    x <- (x - mu.hat) %% (2 * pi)
    x <- matrix(x, ncol = 1)
    z <- apply(x, 1, pvonmises, mu=0, kappa=kappa.hat, tol=1e-020)
    z <- sort(z)
    z.bar <- mean.default(z)
    i <- 1:n
    sum.terms <- (z - (2 * i - 1)/(2 * n))^2
    Value <- sum(sum.terms) - n * (z.bar - 0.5)^2 + 1/(12 * n)
    return(Value)
}

This function is largely inspired from the original watson.test (Circular package). Note the estimate of the two parameters via MLE (the “res” variable). These values are then used to calculate the U2 value:

Now, we just have to define a function that will generate random samples from a true von Mises distribution, and estimate the U2 value for different groups of samples. 


bootstrap.u2<-function(x,n=100){
  samples<-length(x) # number of samples in the original data
  mle.result<-mle.vonmises(x,bias=F) # von Mises MLE parameters estimate for the original data 
  i=0
  
  bootstrap.u2.distrib<-c()
  
  while(i<n){
    boots<-rvonmises(samples,mle.result$mu,mle.result$kappa) # generate deviates from a von Mises with the estimated parameters from the original data
    u2.boots<-get.u2(boots) # get U2 statistics for the current random deviates
    bootstrap.u2.distrib<-c(bootstrap.u2.distrib,u2.boots)
    i=i+1
  }
  results<-sort(bootstrap.u2.distrib)
  return(results)
}

This function takes the original angular values as a vector, calculates the U2 of the original dataset, and estimates the mean and kappa via MLE. Importantly, the estimated parameters values from the original data are then used to generate random von Mises deviates using the rvonmises() function. The number of samples is determined by the original vector size. The U2 is then calculated for these each vector of random deviates, and stored in the bootstrap.u2.distrib vector. This process is iterated for n bootstraps. The function returns an approximation of the U2 distribution for a given kappa and mean, which represents our “null” distribution, that is, the U2 values and their probabilities that corresponds to a given von Mises distribution.

To test whether a given circular distribution is significantly differing from a von Mises, we can count the number of time our U2 statistic from the “null” distribution (returned by our bootstrap.u2() function) falls above the U2 value from our empirical distribution (calculated using get.u2() on our original data). The ‘p-value’ can then be inferred by dividing this value by the total number of U2 samples in the distribution.

References

Lockhart, R. A., & Stephens, M. A. (1985). Tests of fit for the von Mises distribution. Biometrika, 72(3), 647-652. https://doi.org/10.1093/biomet/72.3.647

Sun, Z. (2009). Comparing measures of fit for circular distributions (Doctoral dissertation).
Retrieved from https://dspace.library.uvic.ca/bitstream/handle/1828/2698/zhengsun_master_thesis.pdf

Watson, G. S. (1961). Goodness-of-fit tests on a circle. Biometrika, 48(1/2), 109-114.
https://doi.org/10.1093/biomet/48.1-2.109

Posted on

Machine learning for birds conservation

The ‘microfaune’ (or ‘micro-fauna’ in English) project was started last September within our research division at Wazo, the NGO I co-founded. Microfaune aims to develop machine-learning tools for bioacoustics research, in order to foster birds and wildlife conservation in cities.

The goal of the project is to improve the assessment of urban biodiversity with deep learning algorithms. A first step involves the detection of birdsong from audio recordings, made at the Cité Universitaire de Paris using devices provided by the Cornell University Laboratory of Ornithology. The contributions of this project are:

  • A platform for annotating bird songs (presence or absence)
  • A model allowing the rapid identification chunking and labelling of bird songs
  • An open-source labelled database
The Cornell Lab of Ornithology provided us with state-of-the-art recording tools. Independent recording devices can be installed at specific locations of interest in the canopy. A typical device is able to collect approx 250 GO of data every 15 days.

The project, led by Hadrien Jean and his team, was selected for the Fall 2019 and Fall 2020 season of DataForGood, the French incubator for common good.

The code has been made freely available on github, and can be deployed on Google Cloud AI platform and AWS.

Posted on

Wazo just awarded the Medal of honour from the City of Paris

Wazo – Cité internationale des Oiseaux, a France-based NGO I co-founded two years ago to fight biodiversity decline in cities, has been awarded the prestigious Medal of Honor of the City of Paris. Wazo develops machine-learning algorithms and bioacoustics tools for birds and wildlife conservation. We are particularly involved in certain parks in Paris (e.g. the Cité internationale universtaire park, in the 14th arrondissement).