{"id":"51cd05e6-5ee8-47cb-8c37-e354b64c1c81","arxiv_id":"2412.02846","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Faint broad-line AGN at redshift 4 to 5 appear to live in low-mass galaxies of about 50 million solar masses, implying black holes that are an order of magnitude heavier than their host stellar mass.","lead":"This paper uses the density of neighboring galaxies around faint broad-line active galaxies at redshift 4 to 5 to infer that their host galaxies are about 40 times less massive than standard light-based fits suggest. If correct, these small galaxies host surprisingly heavy black holes, roughly 10% of their host stellar mass.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The untested transfer assumption that BL-Halpha emitters follow the non-AGN overdensity–stellar mass relation is load-bearing; the steep calibration amplifies any environmental offset into a ~0.3 dex shift in inferred host mass.","rationale":"The reader's weakest_assumption correctly identifies the transfer assumption in Section 5 as the key unverified step. I agree: the entire environmental mass inference, and hence the headline BH-to-stellar ratio of 12.5%, depends on BL-Halpha emitters occupying the same overdensity–stellar mass relation as non-broad-line galaxies. The steepness of the calibration (b ≈ 3.2) means a modest environmental bias changes the conclusion by several tenths of a dex. The paper is honest about this assumption and provides useful supporting checks, notably the UV-luminosity-based stellar mass estimate that independently gives 10^7.7 M_sun. That agreement is real evidence, though it shares a separate assumption that the rest-UV light is stellar-dominated rather than scattered AGN light. The BH mass calibration is also uncertain, but the paper explicitly conditions its main claim on 'face value' BH masses, so the least secure novel link is the environmental host-mass inference. A simulation-based test of AGN host environment at fixed stellar mass would directly settle whether the transfer assumption holds; until then, CONDITIONAL is the appropriate verdict.","tokens_in":35573,"tokens_out":15224,"duration_ms":187339,"concrete_test":"Run a controlled simulation test with EAGLE or Illustris-TNG at z = 4.5–5: select galaxies with stellar mass 10^7.5–10^8.2 M_sun, split into an active subset (BH accretion rate or broad-line luminosity matching the ALT threshold, L_broad(Halpha) ≈ 10^43.7 erg/s) and an inactive subset, then measure (1+δ) within 1 cMpc using mock Halpha-selected galaxies with the same luminosity limit and Δz/(1+z) < 0.005. If the mean overdensity of the active subset differs from the inactive subset at fixed stellar mass by more than Δ(1+δ) ≈ 1 (equivalent to ~0.3 dex in stellar mass), the transfer assumption in Section 5 fails and the inferred host mass needs a corresponding correction; if the offsets are smaller, the assumption is validated.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central inference in Section 5 is Equation (1): the overdensity measured within 1 cMpc is converted to a stellar mass through a relation calibrated on galaxies without broad Halpha. The authors state explicitly: 'we assume that BL-Halpha emitters follow the same overdensity - stellar mass relation as galaxies without broad Halpha.' This assumption is the only link between the observable environment and the claimed host stellar mass of 10^7.7 M_sun. The calibration is steep: b = 3.19 ± 0.34 at 1 cMpc, so a systematic offset of only Δ(1+δ) ≈ 1 changes the inferred stellar mass by about 0.3 dex. If AGN activity is triggered by interactions, BL-Halpha hosts at fixed stellar mass could be more clustered than inactive galaxies, biasing the inferred host mass upward; if, instead, broad-line selection favors galaxies in lower-density regions or if the Halpha-selected neighbor census is incomplete around AGN, the inferred mass is too low and the 12.5% BH-to-stellar ratio is overestimated. The companion EW test in Figure 11 shows that the neighbor population is not strongly biased, but it does not test whether the AGN hosts themselves have the same bias as inactive galaxies at fixed stellar mass. The sign and magnitude of this environmental selection effect are unconstrained, and with only six objects the quoted 0.2 dex uncertainty does not include this systematic.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript presents an environment-based method to infer the stellar masses of faint broad-line H-alpha (BL-Halpha) selected AGN at z ~ 4-5, independent of AGN-contaminated SED fitting. Using the deep ALT JWST/NIRCam grism survey behind Abell 2744, the authors measure Mpc-scale overdensities (1 + delta within 1 cMpc cylinders, Delta z/(1+z) < 0.005) around 7 BL-Halpha emitters and around a spectroscopically selected reference sample of 308 galaxies without broad lines whose SED-based masses are well determined. They establish a steep, highly significant correlation between overdensity and stellar mass in the reference sample (Eq. 1: b ~ 3.2, stronger than 6 sigma) and use it to invert the typical BL-Halpha overdensity (mean 5.6 +/- 1.2, excluding the exceptional ALT-66543) into a host stellar mass of log10(Mstar/Msun) = 7.7 +/- 0.2, about 1.5 dex below the median of galaxy-only SED fits. Together with single-epoch H-alpha BH masses (~10^6.8 Msun), this implies a BH-to-stellar mass ratio of ~12.5%. A UV-luminosity-based mass estimate provides a partially independent confirmation. The authors extend overdensity measurements to two z ~ 6.5 LRDs and five EIGER quasars plus one X-ray AGN, finding a tentative correlation between overdensity and BH mass and line width (Fig. 14), and discuss implications for super-Eddington accretion, the AGN duty cycle, feedback-free galaxy formation, and the diversity of AGN hosts.","tokens_in":35823,"tokens_out":21795,"duration_ms":210826,"significance":"The paper's core contribution is a genuinely SED-independent constraint on the host masses of high-redshift broad-line AGN, a population at the center of the 'overly massive black hole' debate; any independent handle on these masses is valuable, and the environment method is clever and, as presented, not circular: the delta-Mstar calibration is fit exclusively to galaxies without broad H-alpha and then applied to BL-Halpha emitters. The paper also ships useful controls: the companion H-alpha EW test (Section 4.5, Fig. 11) addresses the hazard of an H-alpha-selection-dependent neighbor census, and the UV mass-to-light cross-check (Section 5) independently corroborates the low-mass interpretation. The reference sample itself is a strong data product: ~300 galaxies with spectroscopic redshifts and 27-band photometry, with masses spanning 5 x 10^6 to 2 x 10^10 Msun.","major_comments":[{"comment":"The central inference is the conversion of measured overdensities into host stellar masses via Eq. (1), calibrated on galaxies without broad H-alpha. The paper states the assumption that BL-Halpha emitters follow the same overdensity-stellar mass relation as those galaxies, but this assumption is the single most load-bearing element in the analysis and is not tested. Because the calibration slope is steep (b = 3.19 +/- 0.34), a systematic offset of Delta(1+delta) = 1 changes the inferred log Mstar by 0.31 dex, and the quoted 12.5% BH-to-stellar ratio changes by roughly a factor of two. The companion H-alpha EW test in Section 4.5/Figure 11 validates that the neighbor census around AGN is not badly incomplete, but it does not constrain whether the AGN host galaxies themselves are more or less clustered than inactive galaxies at fixed stellar mass (e.g., merger-triggered accretion would bias the inferred mass high; a broad-line selection anticorrelated with environment would bias it low). I ask that the authors (a) quantify the systematic error budget by evaluating the inference under a range of assumed offsets in (1+delta) of, say, +/- 2, and report how the ratio, the comparison with simulations (Fig. 15), and the duty cycle estimate change; and (b) add the closest available direct test: a comparison of the distribution of BL-Halpha overdensities with the distribution for reference galaxies in the 10^7.5-10^8.5 Msun mass bins (data already underlying Fig. 8). I note that the qualitative conclusion, host masses roughly 1 dex below SED-based values, survives plausible offsets, but the quantitative headline numbers do not.","section":"Section 5, Eq. (1)"},{"comment":"The quoted result log10(Mstar/Msun) = 7.7 +/- 0.2 is the scatter among four correlated estimates (1 vs 2 cMpc; all vs centrals), which share the same overdensity data and the same calibration sample, and each of which carries its own reported uncertainty of 0.3-0.5 dex. Propagating the mean overdensity uncertainty (5.6 +/- 1.2) through Eq. (1) alone gives +/- 0.38 dex, and including the errors on a and b raises this to roughly +/- 0.5 dex; the +/- 0.2 dex therefore understates the statistical error by about a factor of two or more before any systematic contribution. The inference is also sensitive to the treatment of ALT-66543: including that object raises the mean (1+delta) to 9.2, which under the linear relation corresponds to log Mstar ~ 8.7, a shift of roughly 1.1 dex (although this point lies beyond the calibrated range, and the authors rightly caution against extrapolation). I ask for a fully propagated uncertainty, a statement of the sensitivity to ALT-66543 in the mass estimate itself (not only in Fig. 12), and the individual inferred masses for the six objects, whose implied values span roughly 10^6-10^9 Msun from the tabulated overdensities.","section":"Section 5, quoted uncertainty"},{"comment":"The title and abstract claim a 'black hole - halo mass relation at z ~ 5,' but Figure 14 combines samples at z = 4.5-6.5 that use different environment tracers (H-alpha for ALT; [OIII] for the UNCOVER and EIGER objects), different luminosity thresholds (L(H-alpha) > 2 x 10^41 vs L([OIII]) > 1.5 x 10^41 vs 10^42 erg/s), and includes two upper limits among the quasars. The apparent correlation is visually driven largely by the two most massive EIGER quasars (J0100+2802 and J0148+0600), which also have the largest Poisson errors. The text appropriately hedges this as 'tentative' and lists the complications, but the title does not. I request either a quantitative censored correlation estimate (e.g., Kendall's tau with upper limits) and a statement of whether the trend survives removing the two most massive quasars, or a tempering of the title and abstract to make the tentative nature explicit. This is a secondary claim relative to the host-mass inference, but it is a headline claim of the paper as titled.","section":"Section 6, Table 3, Fig. 14"}],"minor_comments":[{"comment":"There are two typos: 'at leat' should read 'at least' and 'this studies sample' should read 'this study's sample.'","section":"Section 8, first implication bullet"},{"comment":"The caption says the figure highlights how H-alpha luminosity and width map to SMBH mass and Eddington ratio, but it does not describe how to read the plotted grid contours; please add one sentence of explanation.","section":"Fig. 2 caption"},{"comment":"The sentence 'these measurements do not suggest that BL-Halpha have an excessively large number of nearby pairs' should read 'BL-Halpha emitters,' and the earlier sentence 'do not appear strongly correlated with themselves' should read 'with each other.'","section":"Section 4.2"},{"comment":"Please state explicitly in the figure caption which error bars are shown: the EIGER quasar measurements carry ~0.3 dex systematic uncertainties on MBH that are discussed in the text but not visible in the figure.","section":"Fig. 14 and Table 3"},{"comment":"The UV-based mass estimate of 10^7.7 +/- 0.2 (stat) +/- 0.5 (sys) Msun is presented as corroboration, but it inherits the reference-sample mass-to-light calibration; naming this dependence in the same sentence would make the level of independence of the two estimates clearer.","section":"Section 5, UV cross-check"},{"comment":"The abstract's '~40 times lower' should be quoted as a range (e.g., '10-100 times lower') given the error-budget issues raised in the major comments, to avoid overprecision.","section":"Abstract"},{"comment":"The overdensity-mass calibration is fit in a single ~30 arcmin^2 field; a sentence estimating the possible impact of cosmic variance on the slope and intercept (e.g., from the redshift spikes visible in Fig. 6) would strengthen confidence in the empirical relation.","section":"Section 4.2 / Table 2"}],"recommendation":"major_revision","confidential_remarks":"To the editor: This is a timely and interesting manuscript whose central technique, using environments to infer AGN host masses, is genuinely novel and likely to be influential regardless of the precise outcome. My main reservations are (1) the headline precision (10^7.7 +/- 0.2 Msun) is not a real uncertainty, and (2) the transfer assumption in Section 5, while honestly disclosed, carries the weight of the quantitative claims. I do not think the qualitative conclusion is endangered; the revisions I request are feasible within the manuscript's scope. Please also note the reliance on in-preparation references (Labbe et al. in prep) for the properties of ALT-66543, the object whose exclusion partly drives the headline number."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: this paper introduces a genuinely new way to estimate host stellar masses of faint broad-line AGN—using Mpc-scale overdensities from deep grism spectroscopy—and the punchline, that typical BL-Halpha emitters at z~4-5 live in ~5e7 Msun galaxies, is probably in the right ballpark. But the method rests on an assumption that is stated but untested: that AGN hosts populate the same overdensity–stellar mass relation as non-AGN galaxies. Given the steep calibration slope, that assumption is doing real work.\n\nWhat the paper does well: it builds a clean reference sample of ~300 galaxies with spectroscopic redshifts and good SEDs, demonstrates a strong overdensity–mass correlation, carefully treats satellites, tests the Halpha selection with EW distributions, and cross-checks the result with UV-based masses. The authors are honest about the caveats. The result, if right, shifts the LRD host-mass debate down by ~1.5 dex and poses a sharper challenge to simulations.\n\nThe load-bearing assumption is the one identified in the stress-test note. The authors explicitly assume BL-Halpha emitters follow the same overdensity–mass relation as galaxies without broad lines. The companion EW test shows the neighbor census isn't strongly biased, but it doesn't prove the hosts themselves are unbiased at fixed stellar mass. Since b≈3.2, a systematic offset of ~1 in (1+δ) changes the inferred mass by ~0.3 dex. With six objects, the quoted 0.2 dex scatter among four correlated estimates is not the real error budget. The unpublished companion papers (ALT survey, Labbe in prep, etc.) also make reproducibility hard at the moment. These are limitations, not fatal flaws—the paper frames the result as empirical and conditional.\n\nWho is this for? People working on JWST AGN demographics, SMBH seeding, and simulations of high-z BH growth. It deserves a serious referee; the method is novel and the field needs independent host-mass estimates. I'd suggest the referee push for a quantitative treatment of the systematic due to the transfer assumption, and for the companion data to be posted, but this is not a desk reject.\n\nSend it to review. I would cite it.","headline":"A genuinely new environmental method for AGN host masses that probably points in the right direction, but the central transfer assumption and tiny sample mean the 0.2 dex error is understated.","tokens_in":36481,"tokens_out":2117,"would_cite":true,"duration_ms":22616,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Six faint AGN at z≈5 appear to live in galaxies roughly forty times less massive than their SEDs suggest, implying black-hole-to-stellar mass ratios near ten percent.","keywords":["active galactic nuclei","broad-line H-alpha emitters","black hole–galaxy scaling relations","JWST grism spectroscopy","galaxy overdensity","Little Red Dots","high-redshift supermassive black holes","black hole–halo mass relation"],"falsifier":"Measure the stellar mass of one of these six hosts directly, for example by fitting stellar absorption features in deep JWST/NIRSpec rest-frame optical spectra or by resolving the host galaxy with high-resolution imaging; a value above ~10⁹ solar masses (as the SED fits imply) would contradict the environmental inference, while agreement near $10^{7}$.7 solar masses would confirm it.","tokens_in":35362,"feed_emoji":"🕳️","tokens_out":3383,"duration_ms":41306,"temperature":0.7,"pith_summary":"This paper tries to establish that the faint broad-line H-$\\alpha$ emitters discovered by JWST are hosted by surprisingly low-mass galaxies, around 5×10⁷ solar masses, by measuring how many neighbors each AGN has on megaparsec scales instead of fitting its light. The authors calibrate an overdensity-to-stellar-mass relation using roughly three hundred spectroscopically confirmed star-forming galaxies in the same field, then apply it to six faint AGN with black-hole masses near $10^{6}$.8 solar masses. If the black-hole masses are taken at face value, the inferred hosts imply black holes weighing about 10 to 12.5 percent of their galaxy's stellar mass, far above the local relation. That result would matter because it sharpens the puzzle of how supermassive black holes grew so early, and it points toward either super-Eddington accretion with a low duty cycle or a genuine failure of standard galaxy-formation simulations to grow black holes in small halos.","feed_headline":"Black holes may be 10 percent of their host galaxies' mass","feed_subtitle":"JWST overdensity maps place faint AGN hosts at ~50 million solar masses, 40 times below SED fits.","key_machinery":"The central mechanism is the empirical overdensity–stellar-mass calibration: for galaxies without broad H-alpha, the number of neighbors within a 1 cMpc cylinder (normalized to the random expectation) rises monotonically with stellar mass, with slope b ≈ 3.2 in log10(Mstar/Msun). The calibration is made robust by removing satellite galaxies, defined as systems with a more massive companion within 3 arcseconds, and by checking that H-alpha-based selection does not bias the neighbor counts. The resulting relation is then inverted for the AGN, whose environments are measured with the same prescription, giving host stellar masses independent of the uncertain AGN contribution to the SED.","core_discovery":"The paper's central claim is that the typical faint broad-line H-alpha emitter at z≈4–5 is hosted by a galaxy of stellar mass log10(Mstar/Msun) ≈ 7.7 ± 0.2, about 1.5 dex below the median mass inferred from galaxy-only SED fits. This is obtained by showing that megaparsec-scale overdensity correlates strongly with stellar mass in a reference sample of ~300 galaxies (at >6σ significance), and that the mean overdensity around the faint AGN, 1+δ ≈ 5.6, lands at the low-mass end of that relation. The authors then argue that taking the single-epoch virial black-hole masses at face value yields a black-hole-to-stellar mass ratio of roughly 12.5%, and that combining their measurements with more luminous z≈6 quasars reveals a tentative correlation between line width, black-hole mass, and overdensity, suggestive of a steep black-hole-to-halo mass relation at these redshifts.","pith_inferences":["A direct test would be to measure stellar absorption features (for example Balmer absorption in high-S/N JWST/NIRSpec spectra) in these same six AGN; if stellar masses of order 10^9–10^10 solar masses were found, the environmental calibration would be invalidated for AGN hosts.","The super-Eddington interpretation predicts strong H-alpha variability on rest-frame timescales of months to years, which could be tested with repeated grism observations of the same fields.","If the steep black-hole-to-halo mass relation extends to even fainter AGN, the same environmental technique applied to thousands of sources from upcoming wide-area grism surveys could map the growth of the first black holes without any SED assumptions.","The result adds to the emerging picture that many 'over-massive' black holes at high redshift may be a selection effect amplified by luminosity bias, but it also highlights that low-luminosity AGN in low-mass galaxies are a distinct population that simulations currently fail to reproduce."],"forward_implications":["If the hosts are really as light as ~5×10⁷ solar masses, the implied black-hole-to-stellar mass ratio of ~12.5% is roughly two orders of magnitude above the local relation, confirming earlier 'overly massive black hole' claims with an SED-independent method.","The low inferred host masses place these AGN in strong tension with standard hydrodynamical simulations such as EAGLE and Illustris-TNG, which grow 10^7-solar-mass black holes mainly in galaxies of ~5×10⁹ solar masses.","A duty cycle of ~1% implied by comparing AGN number density to the abundance of low-mass galaxies supports a picture of short, super-Eddington accretion episodes, which would mean the single-epoch black-hole masses may be over-estimated by up to an order of magnitude.","The overdensity results disfavor the alternative that the broad Balmer lines arise from virial broadening in extremely dense stellar systems, since such an interpretation would require host masses near 3×10¹⁰ solar masses with overdensities about three times higher than observed.","The correlation between overdensity and black-hole mass (or line width) implies a luminosity-dependent diversity in AGN hosts, so that deeper surveys will preferentially find AGN in low-mass, blue, unclustered galaxies while rare luminous quasars will be found in massive, red, overdense hosts."],"supporting_citations":[{"why":"Supplies the ALT survey data, the grism sample of galaxies and BL-Hα emitters, and the SED-fitting catalog of stellar masses.","marker":"Naidu & Matthee et al. 2024"},{"why":"Provides the method for identifying and fitting broad Hα components in NIRCam grism data, and the population context of BL-Hα emitters.","marker":"Matthee et al. 2024"},{"why":"Confirms three of the seven AGN spectroscopically with NIRSpec and supplies two z≈6.5 broad-line AGN used in the BH-mass–overdensity correlation.","marker":"Greene et al. 2023"},{"why":"Calibrates the single-epoch virial black-hole mass estimate from Hα line width and luminosity used throughout the paper.","marker":"Reines & Volonteri 2015"},{"why":"Provides the overdensity and halo-mass methodology for z≈6 quasars that the paper extends to fainter AGN, and supplies the comparison quasar sample.","marker":"Eilers et al. 2024"},{"why":"Models the expected luminosity bias in AGN samples and supplies the intrinsic BH–stellar mass relation that the observed sample is compared against.","marker":"Li et al. 2024b"},{"why":"Derives the duty-cycle estimate for BL-Hα emitters from their number densities, which the paper connects to the super-Eddington interpretation.","marker":"Pizzati et al. 2024a"},{"why":"Shows that an EAGLE model without stellar feedback (ONLYAGN) grows SMBHs in lower-mass galaxies, the simulation point of comparison in Figure 15.","marker":"Bower et al. 2017"}],"fun_headline_variants":["Faint AGN hosts 40x lighter than SED fits imply","JWST overdensity maps slash AGN host masses","Black holes reach 10% of host galaxy mass at z~5","Tiny galaxies harbor massive black holes at high redshift"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The method assumes that having a broad H-alpha line does not change how many neighboring galaxies an AGN has at a fixed stellar mass, so that the density–mass relation calibrated on ordinary galaxies applies to the AGN hosts.","fun_headline_variants_meta":{"raw":{"variants":["Faint AGN hosts 40x lighter than SED fits imply","JWST overdensity maps slash AGN host masses","Black holes reach 10% of host galaxy mass at z~5","Tiny galaxies harbor massive black holes at high redshift"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000184,"raw_usage":{"total_tokens":1416,"prompt_tokens":1143,"completion_tokens":273,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":759,"completion_tokens_details":{"reasoning_tokens":202}},"tokens_in":759,"tokens_out":273,"duration_ms":3716,"temperature":1.0,"reasoning_tokens":202,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T23:01:39.715198+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the stellar mass of one of these six hosts directly, for example by fitting stellar absorption features in deep JWST/NIRSpec rest-frame optical spectra or by resolving the host galaxy with high-resolution imaging; a value above ~10⁹ solar masses (as the SED fits imply) would contradict the environmental inference, while agreement near $10^{7}$.7 solar masses would confirm it.","supporting_citations":[],"review_version":1}