June 2017 archive

Jun 29 2017

Book review: Numbers and the making of Us by Caleb Everett

numbersMy next read is Numbers and the Making of Us by Caleb Everett. It is a book about our innate numerical senses, and how we developed skills beyond them that are enabled by the language of numbers.

The book starts with an overview of numbers in writing systems. Highlighting that ways of recording numbers, in the form of tallies, appeared before full blown language. Tallies seem to be evident in prehistoric artefacts before being found in the Fertile Crescent where they were used to record quantities of grain and the like in early written languages. Later in the book Everett proposes that “static” agriculture drives the development of numerical language, citing as evidence the fact that the few remaining hunter-gatherer societies have fewer number-words than their agricultural counterparts. He suggests that as specialisation due to agricultural occurs, it becomes necessary to record amounts of food so that they can be bought and sold (or exchanged for other goods).  

I was interested to read that many human number systems are quinary, decimal or vigesimal, or put more simply based on 5, 10 and 20. This relates to the number of fingers and toes we have, our number words are linked to the numbers of our fingers and toes because they are used for counting. In some languages the link to fingers and toes is explicit, whilst in Western European languages it is not. Vigesimal vestiges remain – in English the number 20 has it’s own special word – score. And in French eighty is quatre vingt and ninety is quatre vingt dix. Later in the book, Everett suggests that our bipedalism helps with the development of counting – our fingers being relatively available (unlike toes) and also important because of our use of tools.

Numbers don’t just come into language in the form of numbers, they can be inferred in the plurals of nouns, or in the forms of verbs. In English we are used to the idea of adding an “s” at the end of a word to make a noun a plural. This is common across many languages. Plurals can also be formed with prefixes (at the start of the word), and even “Infixes”. For example in the Tuwali Ifugao language woman is babai but women is binabai. The plural is formed by adding –in- within the word. But not only this, number can be indicated in the conjugation of forms, and pronoun forms. English is a bit sparse in this regard, other languages have pronouns indicating two or three people as well as just the English divisions of one (I) and many (we).

It seems we have an innate, “exact” number sense for numbers up to three. Beyond three we have a innate, fuzzy number sense meaning we can tell when objects groups are more than or less than one another but where we cannot accurately subject the numbers of objects in a group. Beyond these two senses is a matter of learning.

The evidence for this comes from a number of places, the first of these are experiments with people from anumeric Amazonian tribes and Nicuraguan home signers, who are also anumeric. By anumeric, we mean that they do not have number words beyond three. These people are able to distinguish small numbers exactly but become increasingly inaccurate beyond 3.  

A second strand of evidence comes from the investigation of numeric ability in children, some as young as 48 hours old! Although older children can be questioned directly in spoken language regarding numbers, for younger children it is necessary to use indirect methods. In particular, researchers can track gaze, and watch the absence of sucking on a dummy. Children (and adults), turn their gaze to things they find interesting or unusual and will stop chewing/swallowing/sucking as well. An example experiment in this area is to hide an object behind a screen, and then pretend (or not) to add a second object of the same type, or remove the object. If the child shows excessive gaze, or reduced sucking then it is inferred they are surprised by what they see when the screen is dropped again. That’s to say if they see 2 objects when they expected 1, or vice versa. This surprise implies the ability to count.

Evidence for the exact and fuzzy number sense is also found in experiments on animals. Although some animals, following much training appear to be able to count exactly beyond 3 they are rare. Otherwise they show the same types of innate abilities that we have.

Language is the enabler for exact counting beyond three, clearly sometime in the past one or more humans has learned how to count. Embodying this ability into language enabled it to be transmitted to other, less gifted humans.   

I found this book really fascinating, interested, as I am in both words and numbers.

Jun 05 2017

Book review: Women in Science by Rachel Ignotofsky

women_in_scienceWomen in Science by Rachel Ignotofsky is a whistle-stop tour of 50 women in science mainly from the mid-19th century onwards. Each woman gets a double page spread, with a few paragraphs of text on one page and a cartoon drawing of them and some catchphrases on the other. As well as this there is a centrefold of lab equipment, a timeline and some very brief descriptions of 14 further women in science at the end. You can see more on the authors website, here.

Also included are some statistics on women in science, technology, engineering and maths (STEM), I suspect the figures relate to the US but the picture would not be dramatically different in the UK. On the plus side the proportion of women in STEM has increased from 14% in 1970 to 41% in 2011 and it has been rising steadily. The proportion of engineers who are women rose from 3% in 1970 but has been on a plateau at 13% since 1990. In computer work the proportion of women peaked in 1990 and has been dropping since then, it now stands at 27%.

Why is this important? Historically women have been treated as second class citizens. It wasn’t that they tried to do the things that men did traditionally, and failed. They were very actively prevented from studying in their chosen fields. They weren’t allowed into science labs or science lectures. And if by some chance they did manage to train themselves, there were no jobs or facilities for them to continue their work because they were women. This is the legacy we are trying to overcome.

It isn’t a matter of deep history, women alive today will have been refused access to degree courses in their chosen subjects. Cambridge University, for example, only awarded the first full degree to a woman in 1946, which is the year my mother was born. The parents of men alive today would have kept those systems in place. Women only got the vote in the UK during the lifetime of my grandparents. After I was born my mother was denied an application form for an administrative job at a local garage because the owner felt that her place was at home with her young children. Since the 1970s the spirit of the welfare system in the UK has changed to one in which it is seen as best for both parents to work. And yet historically women have been denied access to many careers. This leaves a legacy because people tend to recruit other people like themselves. The aspirations of children and young people are shaped by the roles they see people like them undertaking.

This book provides a set of role models that show that women can be successful in science.

The 50 chosen women are from a range of sources, many of them are from the rather sparse roll-call of female Nobel Prize winners. Some of the names I recognised: Marie Curie, Jocelyn Bell-Burnell, Jane Goodall, Ada Lovelace, Katherine Johnson (through my very recent reading), Dorothy Hodgkin, Rachel Carson, Lise Meitner. Others I had never heard of, like Lillian Gilbreth who worked on psychology and industrial design. Or Patricia Bath, who founded the American Institute for the Prevention of Blindness.

I’ve looked through the book with my son (aged 5), he seemed to like it – although his main questions on each page were “Where was she born?” and “Where did she go?”. Then again in a book on the history of art his questions were “Where’s Jesus?” and “Why are those people naked?”. I suspect it is better suited to children a little older than him.

Currently my son is binge watching “Horrible Histories”, a programme for children about history. It is a string of vignettes from history acted as adverts, as music videos, game shows or just plain acted. It is lively and educational. It strikes me that Women in Science would provide an excellent source for a sister programme.

I don’t think I am the intended audience for this book but it did remind me to put some more biographies of women in science on my reading list. I’m pleased to see there is a biography of Maria Sibylla Merian, 17th century illustrator and entomologist. Ada Lovelace and Mary Anning are also on my list.

Jun 01 2017

Book Review: Scala for the Impatient by Cay S. Horstmann

scala_for_impatientI thought I should learn a new language, and Scala seemed like a good choice so I got Scala for the Impatient by Cay S. Horstmann.

Scala is a functional programming language which supports object orientation too. I’m attracted to it for a number of reasons. Firstly, I’m using or considering using a number of technologies which are based on Java – such as Elasticsearch, Neo4j and Spark. Although there are bindings to my favoured language, Python, for Spark in particular I feel a second class citizen. Scala, running as it does on the Java Virtual Machine, allows you to import Java functions easily and so gives better access to these systems.

I’m also attracted to Scala because it is rather less verbose than Java. It feels like some of the core aspects of the language ecosystem (like the dependency manager and testing frameworks) have matured rapidly although the range of available libraries is smaller than that of older languages.

Scala for the Impatient gets on with providing details of the language without much preamble. Its working assumption is that you’re somewhat familiar with Java and so concepts are explained relative to Java. I felt like it also made an assumption that you knew about the broad features of the language, since it made some use of forward referencing – where features are used in an example before being explained somewhat later in the book.

I must admit programming in Scala is a bit of a culture shock after Python. Partly because its compiled rather than interpreted, although the environment does what it can to elide this difference – Scala has an REPL (read-evaluate-print-loop) which works in the background by doing a quick compile. This allows you to play around with the language very easily. The second difference is static typing – Scala is a friendly statically typed language in the sense that if you initialise something with a string value then it doesn’t force you to tell it you want this to be a string. But everything does have a very definite type. It follows the modern hipster style of putting the type after the symbol name (i.e var somevariablename: Int = 5 ), as in Go rather than before, as in earlier languages (i.e int somevariablename = 5).

You have to declare new variables as either var or val. Variables (var) are mutable and values (val) are immutable. It strikes me that static typing and this feature should fix half of my programming errors which in a dynamically typed language are usually mis-spelling variable names, changing something as a side effect and putting the wrong type of thing into a variable – usually during I/O.

The book starts with chapters on basic control structures and data types, to classes and objects and collection data types. There are odd chapters on file handling and regular expressions, and also on XML processing which is built into the language, although it does not implement the popular xpath query language for XML. There is also a chapter on the parsing of formal grammars.

I found the chapter on futures and promises fascinating, these are relatively new ways to handle concurrency and parallelism which I hadn’t been exposed to before, I notice they have recently been introduced to Python.

Chapters on type parameters, advanced types and implicit types had me mostly confused although the early parts were straightforward enough. I’d heard of templating classes and data strctures but as someone programming mainly in a dynamically typed languages I hadn’t any call for them. I turns out templating is a whole lot more complicated than I realised!

My favourite chapter was the one on collections – perhaps because I’m a data scientists, and collections are where I put my data. Scala has a rich collection of collections and methods operating on collections. It avoids the abomination of the Python “dictionary” whose members are not ordered, as you might expect. Scala calls such a data structure a HashMap.

It remains to be seen whether reading, once again, chapters on object-oriented programming will result in me writing object-oriented programs. It hasn’t done in the past.

Scala for the Impatient doesn’t really cover the mechanics of installing Scala on your system or the development environment you might use but then such information tends to go stale fast and depends on platform. I will likely write a post on this, since installing Scala and its build tool, sbt, behind a corporate proxy was a small adventure.

Googling for further help I found myself at the Scala Cookbook by Alvin Alexander quite frequently. The definitive reference book for Scala is Programming in Scala by Martin Odersky, Lex Spoon and Bill Venners. Resorting to my now familiar technique of searching the acknowledgements for women working in the area, I found Susan Potter whose website is here.

Scala for the Impatient is well-named, it whistles through the language at a brisk pace, assuming you know how to program. It highlights the differences with Java, and provides you with the vocabulary to find out more.