‘We did not see that our slogans had lost their bearing and pointed in the wrong directions. We invoked “democracy” solemnly as in a prayer, and watched while the greatest nation of Europe voted, by perfectly democratic methods, its assassins into power. We worshipped the will of The Masses, and their will turned out to be death and self-destruction. We regarded capitalism as an outworn system, and were willing to exchange it for a new form of slavery. We preached tolerance, and the evil which we tolerated destroyed our civilization.’
Arthur Koestler, “The Chinaman’s Nod”, in Bricks to Babel.

Bricks to Babel is a collection of Koestler’s writings, many of them short pieces. “The Chinaman’s Nod” is a very short piece, about Koestler’s posting to Berlin as a journalist. Koestler arrived in Berlin on September 14, 1930, the day of the election in which the National Socialists dramatically increased the number of seats they held to 107 of 577, from just 12 seats before. Koestler refers to the day as ‘the beginning of the age of barbarism in Europe’.

“The Chinaman’s Nod” is primarily about a disjuncture between the imagined state of political manoeuvres and political discourse, and the actual behavior of the body politic.

Much of the piece takes the form of a depressing run-down of how the various groups that might have been expected to challenge and arrest the rise of the Nazis – e.g. the Liberals, the Socialists, the intellectuals, the Communists – could not (or did not, until the point it was too late), either because their credibility with the people was poor, or because they were apt to operate in an ineffectual way.

If the disjuncture I described above was the problem, the natural question to ask is about how this disjuncture came to be. Koestler’s answer might refer to the preoccupation of politicians (and others) with issues they thought were important, their failure to recognize real danger, etc. However the underlying pathology he diagnoses isn’t something specific to some subset of citizens; as he writes,

‘After the event, people asked themselves: How could we have been such fools to twiddle our thumbs when the outcome was so obvious? The answer is that owing to the inertia of human imagination, to most people it wasn’t obvious at all.’

I read Koestler’s suggestion of the existence of some sort of ‘inertia of human imagination’ as a reference to something we are all subject to.


What does the talk about political dysfunction look like nowadays? Words like ‘gridlock’ and ‘polarized’ tend to come up when discussing the health of the world’s super-powered democracy. (A body out of balance, and in need of a cleanse, maybe.) People speculate about how the mysterious algorithms behind things like Facebook’s News Feed create political echo chambers for individuals who are subsequently surprised to find that people who don’t share their views actually exist, and in significant numbers. We worry about the fragmentation of discourse, and wonder if it adds up to a coherent imagination of community.

At home, terms like ‘silent majority’ and ‘the 70%’ come up, sometimes (I get the feeling) from people who seem like they feel disappointed or even betrayed that those who disagreed with them failed to do so in terms more to their liking, in media they prefer. At least ours is a small country.

Earlier I asked the question about how the disjuncture between the imagined state of political discourse and who people actually end up voting for comes to be. While I agree with Koestler that it’s often an imaginative gap, when I think about ‘discourse’ and what makes a ‘healthy’ political discourse, one way to think about it is in terms of who gets heard and who gets ignored. There’s also the question of how we hear, which is both a practical question (what do we tend not to hear/read/watch?) and a question of whether we are aware of our cognitive biases (what do we tend not to understand?).

I actually came across the first passage from Koestler I quoted above on the same day I saw news of J.K. Rowling’s defense of Trump’s right to expression on my FB feed. In truth, I don’t think Mr. Trump needs much help on the media air-time front; can he even be silenced? She might say the question is not whether he can be silenced, but whether he ought to be silenced, and I would have to agree. 

But even in that discussion, I’m not sure the principle of freedom of expression should be uniformly invoked to defend the speech of the powerful as compared to the speech of the vulnerable. Really I think the urgent question is, Whom have we not heard, that we need to hear?

The vulnerable always need advocates, I think; but there are also the ‘not-as-loud’, who might not need advocates, exactly, but whom I think of as having a hard time being understood through the noise. Trump’s voice gets represented, but what about his supporters’? In a democracy where these ‘not-as-loud’ might be the majority, the cost of either not hearing them, or hearing them but not understanding them, may be dear.


Wherefore art thou Singlish?

An NYT article by Gwee Li Sui about Singlish has been making the rounds this weekend, following the news from earlier this week about Singlish lexical items being added to the OED.

Being a student of the linguistics of Singlish, as well as something of a Singlish advocate (read: I generally like to talk about Singlish to anyone I think might be interested), both were welcome pieces of news.

But I was also motivated to add some comments of my own, after Haikel’s repost of Alfian’s critical comments, about Alfian’s comments as well as the NYT article.

1. ‘there is no standard Singlish’ (Alfian)

True enough: ‘standard’ is difficult to sanction1; and yet there is something that people recognize to be Singlish when they speak it and hear it.

I know I’m used to the idea that there is a Singlish grammar, and though I would agree that this is an idea deserving of critical scrutiny, I don’t think this is a difficult idea to validate.

Is there a Singlish grammar? I think we do use Singlish systematically, and there are clearly ‘wrong’ and ‘right’ ways to use it (among a given group, at a given time: language is always changing, quickly or slowly, in small ways or large). It is true that our use of it admits a greater range of variation than languages which have been more thoroughly codified, but even when we encounter variation the intuition is that some specific parameters have been varied, rather than suspect that the speaker is speaking not-Singlish, or is speaking Singlish wrongly.

Overall I think it is probably true that ‘there will be differences in the Singlish spoken by a Malay speaker, Chinese speaker or Indian speaker’, and, Alfian contends, ‘mainly in the vocabulary used’. Beyond vocabulary, I think even some grammatical constructions are more frequently used by speakers with different linguistic backgrounds as well, for instance in how indefinite subjects are introduced, e.g. ‘Got one man looking for you just now,’ vs. other ways of expressing the same proposition – but even with this example, it is not the case that a person who wouldn’t use this construction as his/her first choice wouldn’t understand the sentence, or judge it to be ungrammatical by whatever Singlish standards (that word again) they provisionally hold.

To take up the point on vocabulary, is there a Singlish lexicon? Whereas I am personally convinced that Singlish is structurally systematic and would defend the idea, I would be much less eager to attempt to designate the boundaries of a Singlish lexicon. Whether a word is a ‘Singlish’ word, as compared to being simply a Hokkien/Malay/etc. word, or ‘from’ one of these languages, is much more contentious.

I’ve also observed how Malay words like ‘koyak’ and ‘rosak’ end up being appropriate for different things for Singlish speakers who speak Malay, as compared to Singlish speakers who don’t speak Malay (mainly Chinese speakers; I actually collected data about this, based on prompting speakers to make judgments about when these words apply, in response to being shown pictures of dysfunctional objects). Whereas I learned that ‘koyak’ means something like ‘torn’ in Malay, and would apply more to, say, a deteriorating old book or a threadbare shirt, the non-Malay speakers thought it applied best to more solid or mechanical things like a keyboard or a chair.

So I think in general a good question to ask is where are we situated, relative to where our nominally ‘Singlish’ words are from, e.g. the examples of Hokkien that Alfian identified as having been presented as examples of Singlish. Growing up not speaking Hokkien, I would have likewise been prone to identifying Hokkien expressions with Singlish as well, and to that extent I found Gwee’s take on the matter, that ‘Singlish’s status grew so powerful that the Chinese dialects took refuge in it to re-seed themselves’, to ring true.

As concerns the lexicon, I could say either that the Singlish lexicon is capacious, implying that the language has somehow ‘integrated’ the vocabulary and made them ‘native’ to the platform, in some sense; or that it is really quite limited, and Singlish speakers simply accommodate the use of non-Singlish words all the time. Falling somewhere in between, one might say that it is a thoroughly bastardized language, with the latter situation contributing to the reality of the former – which would make it much like English, Singlish’s primary lexifier.

2. ‘Its syntax is drawn partly from Chinese, partly from South Asian languages.’ (Gwee)

In fact, based on the most credible linguistic scholarship (Bao 2005, Sato 2014, some others), I would say that Sinitic and Malay languages had the clearest influences on syntax.

Of the authors I cited, Bao rated the influence of Sinitic languages (Mandarin Chinese, but also Hokkien, Teochew, maybe Cantonese) as being primary, but others (like Sato) contest this and point to the influence of other local non-Sinitic languages.

South Asian languages have definitely provided lexical items that Singlish speakers use, but whereas the influence of Sinitic languages and Malay languages has been demonstrated in the literature, the influence of South Asian languages has not. That said, I can’t rule it out, not having examined the structure of South Asian languages in detail myself. Which brings us to:

2a. ‘any authority on Singlish needs to be fluent in at least four languages so as to avoid any kinds of biases that might arise from his or her own linguistic background’ (Alfian)

I agree with this, as far as the problem is making claims about which substrate languages are the most major influences on the grammar of Singlish, in which case we should have surveyed a good range of potential influences, or refer/defer to relevant sources.

On the other hand, being able to speak a language isn’t necessarily the standard a linguist would hold for being able to understand things about a language; there are many ways to think about languages, and many things about language to be interested in understanding.

The conceit of the linguist who studies the structural organization of language (I would say all linguists rely on this, but to varying degrees) is that we can understand the interrelation of grammatical systems on an abstract, universal level. A syntactician would also probably agree that having such an understanding of a language is materially different from mastering the language, but I think this shows that these are simply different knowledges. It is often the case, for instance, that someone working intensively on Javanese syntax gains some ability to converse in Javanese, but in presenting the work to someone who works on, say, Somali, the fact that the linguist studying Somali doesn’t speak Javanese is not assumed to be an impediment. If there is a difficulty, it would suggest that the data from Javanese either isn’t the most relevant to the proposed problem, or that it’s just poorly analyzed.

In trying to make sense of Singlish – what it is, where it comes from – the concept of a ‘contact language’ comes to mind. I might start an explanation with: Singlish is a language formed in a contact situation. Some of the languages (Sinitic languages, Malay) we regard as being part of the substrate, and the lexifier language (English) is regarded as the superstrate.

In looking at the syntax of Singlish, one gets the sense of an organic hybridity.

When it comes to contact languages, researchers are often interested to observe what features of the substrate end up featuring in the contact language. One situation we refer to is substrate transfer2, which is when a certain grammatical structure or system from the substrate language is reproduced in the contact language.

Another situation we refer to is substrate reinforcement, which is when a certain grammatical structure or system is found in more than one of the major substrate languages. This could have a stabilizing effect (i.e. the feature is more likely to be retained over time), and I would speculate that it probably even promotes transfer in the first place (though some may object to me tying reinforcement with transfer).

A related idea is congruency, which is when a structure is found in both the substrate and the superstrate. This is also thought to facilitate transfer.

When talking about contact languages, the term ‘contact ecology’ tends to come up, and I find that thinking about an idea like substrate reinforcement involves the kind of thinking that tries to discern and evaluate the contribution of multiple factors that give rise to the more-or-less equilibrium state of affairs.

The idea of ‘equilibrium’ also reflects the idea that the situation under examination is fundamentally dynamic: the idea I mentioned earlier about language always changing features in historical linguistics, but is also especially pertinent to contact languages.

From the work I’ve seen about Singlish syntax, I’m convinced that patterns of imperfect reinforcement and levelling give rise to a grammar that is fundamentally hybrid. Alfian cites the example of the deletion of the copular verb (‘is’, ‘are’), which he observes in Malay, but which is also found in Chinese, and which is why I think copula deletion has stuck around. Alfian’s example is, ‘They tired already but still want to play,’ (diorang dah penat tapi masih nak main), and in Chinese, I might say, ‘他们累了还要玩’ (plural-pronoun tired already still want play).

The grammar is also still changing, because some patterns of either reinforcement or congruence leave room for ambiguity3.

TL;DR: How Singlish is spoken does vary, but I think there is a core Singlish grammar. The syntax of Singlish is hybrid.

1. As some contend, the idea of ‘standard’ itself as applied to the languages people speak is always problematic: standards are institutions, but (fortunately) people seem to dislike hewing to just one institution at a time, all the time. 

2. Bao 2005 argues that, for instance, the aspectual system of Chinese can be observed in Singlish, though not every element was perfectly translated. (An example: Perfective aspect is marked by ‘already’ in Singlish, e.g. ‘He eat already.’ In Chinese, I might say ‘他吃了,’ which is, character-by-character, ‘male-pronoun eat perfective-aspect’. We mark tense in English morphologically, e.g. ‘He has eat-en,’ or, ‘He ate,’ and we don’t usually think of aspect as being tied into the grammar of English, but we can indicate aspect with words like ‘earlier’ or with constructions like ‘has finished eating’.)

3. An example: some speakers I consulted thought the sentence, ‘A man wore jacket,’ was not just odd, but outright ungrammatical, while other speakers thought it was perfectly fine. I hypothesized it was about an ambiguity about whether ‘a man’ should be analyzed as a subject or topic, which would affect whether a mechanism from English should be given precedence.


  • BAO ZHIMING (2005). The aspectual system of Singapore English and the systemic substratist explanation. Journal of Linguistics, 41, pp 237-267.
  • YOSUKE SATO (2014). Argument ellipsis in Colloquial Singapore English and the Anti-Agreement Hypothesis . Journal of Linguistics, 50, pp 365-401.

Week 8 (Thursday): Characterizing NPs in the Data

All has not been idle on this front, though I did go through two very uncertain weeks, which were followed by two weeks of reorientating and conceptualizing.

I’m using this blog again to just sort out some thoughts from looking through the data I am compiling. I don’t think I will be able to exhaustively tag everything, but I need to note some general trends and describe how some of the categories are emerging.

Possessive Determiners

At least from a quick first look, possessive determiners (pd) are the most common NPs. Their properties tend to be [+def, +spec]. This preference is really quite evident. Possibly worth comparing to CoSiB corpus – actually may not be too difficult to implement a search, since pd is a closed class. If it is super common, this would be a variational feature of note, that definiteness gets marked this way preferentially, over the article.

Unusual nouns and N-Ellipsis

I haven’t been able to find that many unmarked bare common nouns which are ambiguously (in)definite. Many are mass or kind nouns, or seemingly proper nouns. A possible exception is ‘O’-level, which seems like a proper noun, but part of the case for interpreting it as a common noun being that it can be pluralized – but probably that won’t work. It usually means ‘~ results’ or ‘~ exams’, too, so it’s easier to interpret it as N-ellipsis (nominal ellipsis).


Pro-drop is very prevalent, even in mid-sentence. Comparing mid-sentence Pro-drop with the usual kinds of ellipsis in English, we see that it is usually verbs, modals, their respective phrases, dependent clauses, or entire clauses that are gapped or sluiced – but not NPs.

‘Got’ construction

Some sentences are structured, ‘Got NP?’, and it seems like pro drop applies, i.e. ‘e got NP?’ However, ‘Got NP’ can also be in the topic position, e.g. ‘Got one time they went KL.’ ‘Got’ in the latter construction functions in a way comparable to expletive-‘There’.


Abandoned this post half-way; I decided I probably should not do an exhaustive analysis of trends, since there are too many dimensions and too much data to tag.

Now it’s Week 10, essentially. I’m about 5k words in on the Overleaf document. Some ways to go.

Week 6 (Sunday): NVA

Read some of Baker 2008, and his proposal for a universal three-way distinction between nouns, verbs, and adjectives. Not hard to get behind. Interesting stuff about the unique situations you’ll find adjectives in as well. Makes it easier to talk about lexical items being flexibly recategorized, if there are categories we can assume exist. (Or to rule this out.)

Week 5 (Friday): Pro Drop

It’s been a long time since I’ve been here. It’s definitely not been a good three weeks for the thesis. Some non-thesis things got done, like getting my IPPT cleared, and Chinese New Year.

At the same time I did some peripheral things like making my Honours Thesis document template. I adapted the Carleton College comps project template to our department’s requirements. That’s some time saved somewhere down the line.

I also collected a range of interesting and potentially relevant Singlish utterances on a document sheet in gb4e (so at least I can copy the relevant ones easily to my document). I now have a modest array of data, an array which actually covers quite a nice spread of phenomena.

The elephant in my mental room however, is still in identifying and articulating their nexus of commonality. That feeling of not having made headway is accurate in at least this respect.

Today, I continued crunching through my data list, sorting it and writing some preliminary analyses. In the process I ended up reading about pro-drop, including Sato and Kim’s influential paper on radical pro drop in CSE.

Week 3, Thursday: Contact Situations, Characterizing SG

Mougeon et al. (2005) and De Garcia (2008) about contact situations and linguistic innovation. How best to demonstrate interference (intra-systemic transfer) (Mougeon et al.). What the relationship is between bilingualism and diglossia (De Garcia) (not the same thing, but related).

Bao (2004) on Lexifier Filter and Substrate Transfer, and the uniquely stable linguistic ecology here – insofar as these things are stable. Some challenges to Bao’s analysis based on the data, but the analysis can probably be extended and qualified, since LF and ST are constraints (ranked/violable). Interesting arguments about the consistency of productivity; the role of LF also would apply.