Adrian Zenz (2018): ‘Thoroughly reforming them towards a healthy heart attitude’: China’s political re-education campaign in Xinjiang, Central Asian Survey, DOI: 10.1080/02634937.2018.1507997. Note: all page numbers here are in reference to this, the peer-reviewed article, which I downloaded very legally. Some of the original data (i.e. the specific links to procurement documents in the main table) was replaced with generic links in the Survey article, possibly due to silly style guide issues, which I talk about more below. They can be accessed in the pre-print draft here.
Adrian Zenz is an interesting figure, to say the least. He is very nice to me in my DMs, so I'll admit I did come back and modify somewhat my sardonic little paragraph about his book implying acceptance of gay people like me is a sign of the end of times. Take that for what you will. Regardless, this article is not about him, but about his research, specifically, the paper he published on mass internment in Xinjiang.
This article is the first of many about Xinjiang published by Adrian Zenz. It is from fall of 2018, a time when the Chinese government was still functionally denying the existence of mass reeducation facilities in Xinjiang even while information about them—particularly through Kazakh citizens detained and later released—was making its way into foreign media, especially with reporting by AP, Buzzfeed, and the US-government sponsored Radio Free America, among others.
Zenz seeks to show that despite denials by the Chinese government, there existed ample evidence from contracting documents that complemented firsthand testimony and demonstrated the reality of a huge surge in the construction of securitized facilities designed for large-scale, forced detention of Muslims in Xinjiang. In his own words:
This article demonstrates that one can also ﬁnd a substantial body of governmental sources, produced for a domestic audience, that prove the existence of the camps. Oﬃcial public tenders indicate the construction of such facilities on city, county, township and village levels. Based on the available documentary evidence, we can surmise that the region’s current re-education system exceeds the size of China’s entire former ‘education through labour’ system, which was oﬃcially abolished in 2013. (p. 2, emphasis added)
What are procurement/construction bids, and what do they prove?
Bids are basically contracts posted by governments to solicit companies to provide a service or good, such as building of a certain facility or procurement of certain equipment. They're a common tool of functionally all modern governments, which, at the end of the day, are not construction or IT firms and need outside expertise to implement policy goals. Zenz gathered bids from online platforms designed to connect Chinese firms with government bodies that put out these contracts (the US operates a similar platform at https://beta.sam.gov/).
What Zenz found was that beginning in mid-2017, Xinjiang saw a massive spike in bids calling for the construction and outfitting of highly securitized detention centers all across the region. These bids were specifically labelled with terms such as 'legal system training' and 'transformation through education,' terms that have become hallmark euphemisms for the mass internment and reeducation program. This is confirmed by Qiu 2017, an article published by Xinjiang Party School researcher Qiu Yuanyuan and cited by Zenz. (I did a partial translation here.)
Here are some selected translations and commentary on a few of the bid documents:
What we see overall is ultimately that beginning in late 2016, and greatly accelerating in 2017, the local governments across Xinjiang began massive expansion of reeducation centers in accordance with policy directives. These centers were coercive, as is very clear from the language of detention and surveillance used and the infrastructure built per bid documents, including large perimeter walls, barbed wire, and surveillance equipment. What exactly goes on inside the camps, however, is not revealed by bid documents—for that, we need to turn to witness/detainee testimony, leaked documents, and other sources. (Zenz' article details this somewhat, citing testimony primarily relayed by US-government backed media sources; assessment of such testimony will be the subject of another post.)
What else does this article cover?
The article is not just about bidding documents, though that is the main portion I have focused on here. It also cites a large number of other sources of varying reliability to contextualize the construction projects within the larger scope of the Communist Party's treatment of Uyghurs. I encourage you to read the paper for yourself. Several of the sources cited are in Chinese, and of those I considered more enlightening, I have provided translations. These include:
- A translation of Turpan County's 2013 documented efforts targeting Muslim religious practices
- The 2017 Xinjiang De-extremification Regulations, which are frankly rife with extremely concerning and Islamophobic stipulations
- The aforementioned published academic research of a Xinjiang Party School researcher, Qiu Yuanyuan, on mass internment and reeducation in its early stages
Methodological Problems and Reproducibility
The final, peer-reviewed edition of this article as it appears in Central Asia Survey does not include full URLs for the procurement bids, which is perhaps a (perplexing) stylistic decision. The pre-print version of the article, available on SocArXiv, does include the table with full URLs as an appendix. Several citations do not appear to have been archived, inexplicably, and some are search engine links (like to Sougou) that should not have been used as citation URLs in the first place. This strongly suggests a lack of familiarity with Chinese internet media, and is highly perplexing: anyone who works with sources on the Chinese side of the Great Fire Wall knows how rampant link rot is, not to mention deliberate deletion, and it is entirely unclear why some sources would be archived while others not.
This necessarily impinged the overall credibility and rigor of the argument Zenz makes. To that end, I contacted him, and he gladly provided a link to download his supporting documentation and explained he was not very familiar with archiving, preferring instead to save everything manually. Fair enough. With with the additional data Zenz shared—available here, a Dropbox file of saved web files that you'll have to download and open in a web browser to read properly— I ultimately consider the problem moot.
Attached below is the appendix extracted and with the reproduction results for each and every contract. I'll be real with you: I did not read through every single word of every single contract, some of which are extremely long and extremely boring. I did at least try to spend a minute or so on each and verify certain factual descriptions (cost, facilities, etc.) in the article, and noted a few minor discrepancies.
In the above, green indicates I was able to access and examine at least one of the source URLs, which I checked briefly against Zenz' descriptions. Red indicates I was unable to do so (all bare/unarchived URLs were also run through the Internet Archive's WayBack Machine manually). Orange indicates I was able to pull up an archive that showed the existence of a project named by Zenz, but the details described by him were not fully accessible. My initial search showed that almost half of the links were dead or otherwise unavailable, which is an extremely low number. After messaging Zenz on Twitter, he quickly provided me a copy of his saved files he maintained on Dropbox. The purple highlights are those that were initially unverifiable but I was able to confirm with the data Zenz sent. Overall, of the 78 bids listed in the article, 39 are immediately verifiable, 33 are verifiable through additional saved data provided by Zenz, 2 are partially verifiable, and 4 are not verifiable.
On a final note, Zenz' Chinese was shaky, it seems. For example, when citing a primary source, he mistakes a village name as Zhudunbage 驻墩巴格村, when the first word, 驻, is actually a verb, 'to be stationed in'. In another example—in a preprint only, as Survey did not include characters in inline parentheticals—保留所 ‘detention center' was misspelled as 抱留所. I was unable to locate an instance where mistranslation impact the arguments made within the paper, however.
This paper is far from perfect, and Zenz is far from a perfect messenger. The core research, however, is solid. Especially combined with the wealth of testimony now available, extensive satellite imagery of camps, and documents leaked to foreign media (some of which I will begin summarizing in the coming weeks), the article is a fairly strong indictment of the current system of mass internment and reeducation of Uyghurs and other Muslim ethnic minorities in Xinjiang.
A month or so after the publication of this article, the Chinese government evidently realized its current strategy of denial was no longer sustainable. Shohrat Zakir, the governor of Xinjiang, gave an interview in Xinhua that described the camps and defended their purpose. From then on, the focus of the Chinese government has been portraying the camps as moderate, humane, and effective, rather than denying their existence. If these camps were so patently benevolent, one does wonder why they were so vigorously denied by the government for so long.