Parsing puzzle of the week

« previous post | next post »

"Short Wave: A Physics Legend", NPR Up First 4/3/2022 [emphasis added]:

In the 1950's, a particle physicist made a landmark discovery that changed what we thought we knew about how our universe operates. Chien-Shiung Wu did it while raising a family and an ocean away from her relatives in China. In this episode from NPR's daily science podcast Short Wave, we delve into the life and impact of Chien-Shiung Wu, widely considered the "queen of nuclear physics."

The bolded sentence is a good example of why parsing can be hard. What's the scope of "and" in "while raising a family and an ocean away from her relatives"? On purely formal grounds, the most obvious answer is the coordination of "a family and an ocean" — which is how Spacy construes it:

Dep tree          Token     Dep type Lemma    Part of Sp
───────────────── ───────── ──────── ──────── ──────────
            ┌───► Chien     compound Chien    PROPN     
            │┌──► -         punct    -        PUNCT     
            ││┌─► Shiung    compound Shiung   PROPN     
         ┌─►└┴┴── Wu        nsubj    Wu       PROPN     
┌┬───────┴────┬── did       ROOT     do       AUX       
││            └─► it        dobj     -PRON-   PRON      
││            ┌─► while     mark     while    SCONJ     
│└─►┌┬──┬─────┴── raising   advcl    raise    VERB      
│   ││  │     ┌─► a         det      a        DET       
│   ││  └─►┌──┼── family    dobj     family   NOUN      
│   ││     │  └─► and       cc       and      CCONJ     
│   ││     │  ┌─► an        det      an       DET       
│   ││     └─►└── ocean     conj     ocean    NOUN      
│   │└─►┌──────── away      advmod   away     ADV       
│   │   └─►┌───── from      prep     from     ADP       
│   │      │  ┌─► her       poss     -PRON-   DET       
│   │      └─►└── relatives pobj     relative NOUN      
│   └────────►┌── in        prep     in       ADP       
│             └─► China     pobj     China    PROPN     
└───────────────► .         punct    .        PUNCT     

Spacy also concludes that she did the raising in China, away from her (other?) relatives (wherever they were)…

The Berkeley Neural Parser does essentially the same thing, though placing the relatives (rather than the raising) in China:

Given the sequence of potential post-modifiers  "raising A and B away from C in D", there are quite a few alternatives. Perhaps someone in the comments will count them :-)…

 



17 Comments

  1. Gregory Kusnick said,

    April 9, 2022 @ 2:07 pm

    If what we thought we knew about the universe was wrong, then maybe raising an ocean isn't out of the question.

  2. Jamie said,

    April 9, 2022 @ 4:00 pm

    I read it as: while (raising a family) and ([while being] an ocean away from her relatives in China). So, very differently from Spacey.

    It seems almost like an example of zeugma.

    [(myl) That's pretty clearly what was intended.]

  3. Brett said,

    April 9, 2022 @ 7:09 pm

    Chien-Shiung is usually referred to as "Madame Wu" to many physicists today—including myself. I think this is partly due to a longstanding uncertainty about how to pronounce her first name correctly, compounded with (in my case at least) a desire not to be disrespectful to a scientist of such stature. (She could easily have shared the Nobel Prize for the discovery of parity violation with Lee and Yang; but she didn't, nor was she the only female physicist to get rooked out of a Nobel.)

  4. David L said,

    April 9, 2022 @ 7:17 pm

    'Raising the ocean' reminds me of a remark attributed to Lew Grade, a tv and movie producer in Britain. He made a film 'Raise the Titanic,' which was an expensive flop. He was reported to have said afterwards, 'it would have been cheaper to lower the Atlantic.'

  5. maidhc said,

    April 10, 2022 @ 4:10 am

    Don't Raise the Bridge, Lower the River

    There was a time when every British movie had Terry-Thomas in it.

    Usually wearing a British Warm.

  6. Cervantes said,

    April 10, 2022 @ 7:32 am

    This is of course no problem at all for humans because we have semantic understanding and knowledge of the world. You actually have to be on the lookout for such a problem before you would even notice it.

  7. Timothy Rowe said,

    April 10, 2022 @ 2:42 pm

    The sentence I found it most fun to parse was from Virginia Woolf's "To the Lighthouse". "Since he belonged, even at the age of six, to that great clan which cannot keep this feeling separate from that, but must let future prospects, with their joys and sorrows, cloud what is actually at hand, since to such people even in earliest childhood any turn in the wheel of sensation has the power to crystallise and transfix the moment upon which its gloom or radiance rests, James Ramsay, sitting on the floor cutting out pictures from the illustrated catalogue of the Army and Navy stores, endowed the picture of a refrigerator, as his mother spoke, with heavenly bliss. It was fringed with joy."

    I think it can be parsed in a couple of different ways (and I suspect the Berkeley Neural Parser is led astray by 1920s British punctuation conventions not being the same as current US ones).

  8. John Swindle said,

    April 10, 2022 @ 3:18 pm

    At first I thought she was raising a family and possibly an ocean here, for some “here,” but her relatives were in China. It seemed odd to distinguish family from relatives so sharply. Then I realized that there was no implication that all her relatives were in China.

  9. Philip Taylor said,

    April 11, 2022 @ 5:29 am

    "It seemed odd to distinguish family from relatives so sharply" — it did not seem odd to me. For example, my sister's family live in Sài Gòn, whilst her relatives live in Đà Nẵng and other places. For me, once one is married, family = spouse + children, relatives = aunts, uncles, etc. Parents are both, of course. but primarily family.

  10. John Swindle said,

    April 11, 2022 @ 8:08 am

    Whereas for me the terms "family" and "relatives" stretch a little, making them either inconveniently imprecise or conveniently inclusive. But it's quite possible to take "her relatives in China" to mean "those of her relatives who were in China," which I decided was preferable to my pretending at cultural analysis!

  11. Daniel Barkalow said,

    April 11, 2022 @ 11:11 am

    This seems like a great example of why you really want to be able to push back on your parser, and also why you want alternatives for plausible structures. From a semantic point of view, it's very unlikely that she was raising an ocean, and even if she was, it would be unlikely that she was also raising a family in the same sense. You want to be able to reject a particular link and get a new parse that's more plausible. (For that matter, you'll probably also need to reject a parse in which she's an ocean before getting one that "an ocean" modifying "away".) But then you'll end up with a situation where "in China" could either be where her relatives are or where she is, and you want to understand the sentence as saying that either she's in China or she's an ocean away from China.

  12. Philip Anderson said,

    April 11, 2022 @ 2:01 pm

    While X’s family can be ambiguous, depending on X’s role in that family (as head or junior member), in the context of raising a family it can only refer to her children. Her relatives then refers to other (adult) relatives, probably excluding in-laws, who would normally help

  13. John Swindle said,

    April 12, 2022 @ 1:43 am

    Indeed, her raising a family can only refer to children, although I suppose they could be her grandchildren or something; and in the context of childrearing the reference to relatives in China does make sense.

  14. Coby Lubliner said,

    April 12, 2022 @ 12:47 pm

    Brett: I was a grad student at Columbia when the physics Nobel Prize was awarded to Lee and Yang, but, scandalously, not to Wu, who had done the experimental work underlying their theory. I remember her always being referred to as C. S. Wu.

  15. Philip Taylor said,

    April 12, 2022 @ 1:54 pm

    [family v. relatives] Radio 4 today, The Caretakers, Episode 2, delta-T = 15:56++ (in translation, original can be heard behind) — "That's really important for me, since I don't have a family here; my father is living in Colombia, so I only have my children and my husband".

  16. Philip Anderson said,

    April 12, 2022 @ 7:50 pm

    @Philip Taylor
    I would have expected the speaker to have said “since I don’t have family here”, rather than “a family”.

  17. Andrew Usher said,

    April 13, 2022 @ 6:57 am

    Yes, that was translated, so not wholly idiomatic English. The subtleties of 'family' may be missed easily, though native speakers recognise them.

    To the original quote I say that it's basically a mistake, though as usual the meaning is clear, putting a comma after 'family' would have been an improvement.

    Coby Lubliner:

    While I don't care that much for the Nobel Prize, it seems to me that in science theoretical understanding is always more important than the experiment work, right? So it's hardly surprising if the Prize is awarded on that basis.

    k_over_hbarc at yahoo dot com

RSS feed for comments on this post