Lost in Translation?

The names of the translators of books written in a foreign tongue, and critical acclaim for their work, are rarely seen on the covers or title pages of English editions published in the UK or the USA. I learnt from Jennifer Croft, a translator writing for The Guardian, that although the Man Booker International Award decided in 2016 forthwith to split the £50,000 prize between author and translator, nevertheless, not one of the six winning works of fiction published since then has the translator’s name on the front.

Translation is a highly skilled art. It calls for a deep knowledge of both languages and a sensitive ear for literary style, social context, idiom and narrative mood. Take the first page (in the original) of your favourite foreign-language novel, cut and paste it into Google Translate, and compare what Google suggests with the first page of an accredited English translation. Convinced?

Jennifer Croft translates into English from Polish, Argentine-Spanish, and Ukrainian and is perhaps best known for translating the Polish novelist Olga Tokarczuk's Flights, mixed-genre work for which Croft and Tokarczuk won the above-mentioned prize in 2018. As she says: "Generally speaking we are also the most reliable advocates for our books, and we take better care of them than anybody else. Covers simply can’t continue to conceal who we are. It’s bad business, it doesn’t hold us accountable for our choices, and in its wilful obfuscation it is a practice that is disrespectful not only to us, but to readers as well."

The urge to correct

Scientists who analyse large amounts of data in their research often store it in Excel spreadsheets. Excel has for decades been a standard component of the suite of applications in Microsoft Office, but, despite regular upgrades, it has not lost the irritating urge to correct what it often dumbly supposes is one of your errors. Although it's not too difficult to turn off autocorrect, not everyone knows how to or can be bothered to find out. Also, in normal circumstances it can be a helpful tool, provided that you carefully read through what you thought you had written before you launch it into the public domain. (And let's be grateful that Clippy was put back in the box.)

It seems that autocorrect is a particular problem for geneticists. A gene called Membrane Associated Ring-CH-type finger 1, commonly known as MARCH1 for short, is, for instance, frequently corrected to the date March 1st. Something similar happens to genes known as SEPT1 and  DEC2 and there are many other examples both in English and other languages.

Although the problem was first noticed in 2004, it wasn't until 2016 that Mark Ziemann drew wider attention to this hazard. Then, last July he and some colleagues published a paper in the open source journal Public Library of Science (PLOS) Computational Biology entitled “Gene name errors: Lessons not learned". He and his co-authors surveyed 166,000 genomics-related papers published between 2014 and 2020, and they found that the number of papers using Excel had steadily increased and the proportion plagued with autocorrect errors still hovered around 30%. Various remedies have been proposed: a change in the official names of genes to make them less tempting targets for correction, the use of bespoke scientific software for data processing, ...  and (my naive idea) to make the Excel default "autocorrect OFF". (My thanks to The Economist of 1st September 2021 for telling me about this.)

Sonnet for Silenced Singers

 “Bare Ruin’d Choirs”

The church hall sleeps in evening silence now

The singers’ cars no longer line the lane

The maestro’s voice that rucked the tenor’s brow

No longer snaps “Repeat that last refrain”.

A speckless microbe felled the human giant

Who’s now confined to languish in his den.

With Government instructions e’er compliant

No consort, choir or chorus sings Amen.

From data coming in they’re getting wiser,

And lab-coats warn the virus “We’ll subdue ya”.

With jabs from AstraZeneca and Pfizer,

They can now let rip again with “Alleluia!”

    Their joyful choral cadences restored,

    In four-part harmony they’ll “Praise the Lord”.

Graduate longevity: cause or correlation?

The year of my birth was a perfect square, and y years ago my age was the fourth power of y. I started university when I was 21,  There were 143 in the cohort of undergraduates starting that year at Christ's College, Cambridge. An actuarial friend has kindly estimated that only around 32% of males born in Britain in the same year as me are still with us; taking that as the average should mean just (32 x 143)/100 ≃ 47 of the Christ's cohort remain among the quick. In fact, 83 of them are still alive and more or less kicking - 36 more than would be statistically expected if longevity were distributed evenly through society. If you wish to live longer, it seems you should get a degree, or perhaps partake of the waters and the wine at Christ's.

A Ton

This is my hundredth post (although from my laptop’s point of view it’s post number 1100100 since 100 = 64 + 32 + 4).

... and Statistics

The number 360,960 looks frighteningly large. It’s the headline-grabbing number of new confirmed cases of covid-19 in India on 22nd April 2021. It was thus reported in The Guardian, a UK newspaper. as “a world record for any single country since the pandemic began”. By the time you read this the daily total may even be over 400,000 but that doesn’t significantly change the comparison we make below.

These sensational headlines are as dramatic as they are dishonest: They convey the false idea that India’s a basket case when it comes to daily covid figures getting out of control. 

But let’s look more closely. On 9th January 2021 the United Kingdom reached a peak of 89.3 new daily cases per 100 thousand. The corresponding figure in India on 22nd April was 26.4 (or 29.3 if you use the 400K figure). In either case the comparable daily rate in the UK was over 3 times worse than India’s, and England’s alone was even higher. Since the population of India is over 20 times greater than the UK’s, it’s hardly surprising that the country clocks up an alarmingly huge grand total even when it’s doing quite well on a daily basis compared with other countries.

The more honest headline might have been: “India’s medical infrastructure struggles to cope with a comparatively modest daily rate of covid-19 cases.” and this similarly serious but less eye-catching observation should also raise no eyebrows given how poor India is in per capita terms. Unfortunately the sober truth doesn’t keep the presses rolling and the anchors watched.

The Long and the Short of it

One of our sons (D) and his partner (L) had covid-19 at the start of the pandemic in March 2020. They both had the standard covid symptoms but there were no available tests to confirm the diagnosis at that time. D was off-colour for all of three days with symptoms no worse than a cold,  whereas L, our de facto daughter-in-law, has been laid low for over a year and is still suffering the debilitating consequences.

Like L, many long-covid sufferers have not had a positive test for SARS-CoV-2 virus or for antibodies against it, either because tests were not available when they were ill or because  the antibody tests they had were not sensitive enough (a problem with several of the early tests). 

Broadly speaking, there are three types of long-covid symptoms: 

1. The first type is characterised by “exercise intolerance”, whereby the sufferers have little stamina and get exhausted from even the smallest tasks involving physical activity.
2. The second type involves cognitive disturbances, such as a sensation of ‘brain fog’ and memory problems. 
3. The third type affects the autonomic nervous system, which controls things like heartbeat, breathing and digestion. Patients in this group can experience heart palpitations and dizziness for example.

L has endured an extreme case of type 1 symptoms, as well as some symptoms of type 2. There are some first-hand accounts from sufferers in this Royal Society video.

Medical researchers are exploring three possible biological explanations:

Hypothesis 1. Long covid is a lingering viral infection which can hide out of reach of the immune system (like malaria, chickenpox and HIV), perhaps producing some active viral proteins not detectable by PCR and other antigen tests.

Hypothesis 2. It is an auto-immune disorder. There is strong evidence of cases of an over-reaction of the immune system in the acute stage of covid when many patients died from so-called cytokine storms

Hypothesis 3. It is due to tissue damage caused by inflammation during the initial acute stage, damage for example to the lining of blood vessels restricting blood flow to the brain.

To date there have been almost 4.5m confirmed cases of covid-19 in the United Kingdom and in reality probably many more. The Zoe Covid Symptom Study based in King’s College London found that 1 in 50 covid sufferers still had symptoms 12 weeks or more after the acute phase. This means that getting on for 100,000 UK citizens have post-covid-19 syndrome, and that many of them will need long-term treatment and possibly be unable to work. These cases will place a significant burden on the economy and the medical services over and above the existing negative impacts of the acute covid-19 cases. For this reason, there is hope that more research funding will be channelled towards research into long-covid in particular and post-viral illnesses in general, attracting more medical scientists to direct their efforts to understanding these conditions and finding treatments. That could bring relief and credibility to the many others who have found themselves languishing in similar ways after other viral or bacterial infections. “It’s an ill wind …”

(This blog post is informed by an excellent three-page article entitled “Post-covid syndrome — The sting in the tail” published in the print edition of The Economist on 1st May 2021, page 69.)