Chris Brew [email protected]

natural language processing scientist, data intensive linguist, professor, textbook author

[Chris Brew [email protected] @cbrew]

@40PercentGerman Nothing to do with grammar. It’s perfectly good grammar to use the “wrong” one, just rude.

[Chris Brew [email protected] @cbrew]

@skdh In spoken language, German modal particles translate into variations in intonation. ‘Doch’ translates into the marked intonation contour in “you CAN do that” . You need something to put the intonation in, of course, so it is not as neat.

[Chris Brew [email protected] @cbrew]

@davidfrum Have you heard of Joschka Fischer? He was a violent radical, then Foreign Minister of Germany. I guess you must think that is not a leadership position.

[Chris Brew [email protected] @cbrew]

@TheSpinDr7 @GadSaad I don’t think my Western compatriots are serious about their culture or identity, but they are very serious about the privilege and power that has come with them for the last 300 years.

[Irish Ranger (Sevvy) @VeteranIrish]

British Army Humour

An innocent typo was made and all parties involved have gotten as much enjoyment as they could out of it!

Lt. Colonel Robert Maclaren retired from the British Army in 2001 after a long fulfilling career. On the day that he retired he received a letter from the Personnel Department of the Ministry of Defence setting out details of his pension and, in particular, the tax-free ‘lump sum’ award, (based upon completed years of service), that he would receive in addition to his monthly pension.

The letter read,

"Dear Lt. Colonel Maclaren,
We write to confirm that you retired from the Royal Scots Dragoon Guards on 1st March 2001 at the rank of Lt Colonel, having been commissioned into the British Army at Edinburgh Castle as a 2nd Lieutenant on 1st February 1366.
Accordingly your lump sum payment, based on years served, has been calculated as £68,500. You will receive a cheque for this amount in due course.
Yours sincerely
Army Paymaster”

Colonel Maclaren replied;

“Dear Paymaster,
Thank you for your recent letter confirming that I served as an officer in the Royal Scots Dragoon Guards between 1st February 1366 and 1st March 2001 – a total period of 635 years and 1 month.
I note however that you have calculated my lump sum to be £68, 500, which seems to be considerably less than it should be bearing in mind my length of service since I received my commission from King Edward III.
By my calculation, allowing for interest payments and currency fluctuations, my lump sum should actually be £6, 427, 586, 619. 47p.
I look forward to receiving a cheque for this amount in due course.
Yours sincerely,
Robert Maclaren (Lt Col Retd)”

A month passed by and then in early April, a stout manilla envelope from the Ministry of Defence in Edinburgh dropped through Col. Maclaren’s letter box, it read:

“Dear Lt Colonel Maclaren,
We have reviewed the circumstances of your case as outlined in your recent letter to us dated 8th March inst. We do indeed confirm that you were commissioned into the Royal Scots Dragoon Guards by King Edward III at Edinburgh Castle on 1st February 1366, and that you served continuously for the following 635 years and 1 month.
We have re-calculated your pension and have pleasure in confirming that the lump sum payment due to you is indeed £6, 427, 586, 619. 47p.
However, we also note that according to our records you are the only surviving officer who had command responsibility during the following campaigns and battles;
The Wars of the Roses 1455 -1485 (Including the battles of Bosworth Field, Barnet and Towton) The Civil War 1642 -1651 (Including the battles Edge Hill, Naseby and the conquest of Ireland) The Napoleonic War 1803 – 1815 (including the battle of Waterloo and the Peninsular War) The Crimean War (1853 – 1856) (including the battle of Sevastopol and the Charge of the Light Brigade) The Boer War (1899 -1902) World War One (1914-1918).
We would therefore wish to know what happened to the following, which do not appear to have been returned to Stores by you on completion of operations:
9765 Cannons
26,785 Swords
12,889 Pikes
127,345 Rifles (with bayonets)
28,987 horses (fully kitted
Plus three complete marching bands with instruments and banners.
We have calculated the total cost of these items and they amount to £6,427,518.119.47p.
We have therefore subtracted this sum from your lump sum, leaving a residual amount of £68,500, for which you will receive a cheque in due course.
Yours sincerely . . . .”
😁😁😁

[Chris Brew [email protected] @cbrew]

@JujuAadi @AravSrinivas Odd view of simplicity. With the right database your complex one is easy SQL. Your simple one requires law school or equivalent to do well.

[Chris Brew [email protected] @cbrew]

@alexandersclark This view pretty definitely shaped by early exposure to dynamic logics as arbitrarily general semantic formalisms. You can do anything, but it doesn’t mean you should.

[Chris Brew [email protected] @cbrew]

@alexandersclark I don’t think one can ever say that a natural language is non-compositional, because there is always the alternate possibility that it is compositional but you haven’t yet found the right decomposition into parts.

[Chris Brew [email protected] @cbrew]

@ProfMaxNew If you are talking of NL parsing the issue is that parsing generates a representation that most obvious use cases don’t need. They can jump straight to downstream products like (heavy quotes) ‘meaning’

[Chris Brew [email protected] @cbrew]

@AlwaysUhhJustin @random_walker I’ve also talked to the Microsoft people who support company deployments of 3.5. They haven’t been told much about what OpenAI’s plans are, apparently. The two companies are not (yet??) sharing all the detail with each other. Maybe a result of OpenAI’s charter.

[Chris Brew [email protected] @cbrew]

@matbury @adelegoldberg1 It would certainly fail formal language theory.

[Chris Brew [email protected] @cbrew]

@DCCyclone @mixedlinguist @nytimes I think I agree that he deserved to be mocked. What he did was to unintentionally reveal how he was thinking in trying and failing to get it right.

[Chris Brew [email protected] @cbrew]

@stephlepp Readers, deprived of the easy signals that used to indicate that a piece is not worth their time, may not have the patience to read deeper and notice the difference between “good” and ‘looks good”. Expect lots of fake David Foster Wallace.

[Chris Brew [email protected] @cbrew]

@stephlepp No. But I am assuming that input by humans who have something to say is needed to produce good work. I’m sure we’ll see excellent stuff done by people using AI models to augment what they do.

[Chris Brew [email protected] @cbrew]

@mixedlinguist @nytimes Yep. And not everyone accepts the standard linguist view that some of this is just arbitrary. When we are deep in it emotionally, that view doesn’t feel right, even if it is correct, which I think it is.

[Chris Brew [email protected] @cbrew]

@sitKonrad @emilymbender @Sonja_Drimmer Yes. Some of my colleagues have worked out ways of getting it to provide checkable citations. Sometimes. But ad hoc, messy and untransparent

[Chris Brew [email protected] @cbrew]

@betsysneller The chicken is ready to eat. Ambiguous, though might not be syntax, could be semantics or even semantax . Carl Pollard’s word for the between thing of syntax and semantics.

[Chris Brew [email protected] @cbrew]

@compositiomal Plausible claim. Some syntacticians don’t believe in movement. Do they do the gesture?